Nine months after launching Uphoff Advisory and Uphoff on Media, I had a discovery problem. Seventy-four posts, a growing framework library and a suite of client tools, and not enough of it was being found in search or recommended by AI.

Here’s the admission: I’d spent 20-plus years running companies with teams who did this work. I knew what good looked like. I didn’t know how to do it.

So I handed it to my agents. This was my own back office, not client work. They built or rebuilt 27 framework pages, re-rendered about 30 charts and audited all 74 posts for broken links.

Working alone, a skilled person would need about four weeks to do that. I spent nine hours.

It’s tempting to say AI did the work. It didn’t, and a lot of operators are building their plans on that mistake. The agent didn’t take me out of the work. It moved me from production to judgment. The leverage came from that shift, not from the automation.

Subscribe now

Solo Scale was a thesis. The data is starting to come in

In July I argued in Solo Scale that agentic AI breaks the constraint that has kept every services business at services margins: one expert, finite hours, revenue that scales with headcount. I wrote that the operator “sets strategy, reviews output, exercises judgment” while agents handle the production.

That was an argument. This is a test of it against my own business, with the limits stated up front. It’s one project and one operator. The counts of pages, posts, links and URLs are exact. The hours are estimates, not time sheets, and nine is the high end of mine. Read them as an order of magnitude, not an exact count. The four-week figure is a line-item estimate, not a guess: writing 27 pages is the biggest piece at 40 to 68 hours, followed by building them in WordPress and re-rendering the charts.

Four weeks against nine hours is roughly 18 to 1. At the low end of the estimate, 125 hours, it’s still 14 to 1.

Where my hours actually went

The division of labor was clean on paper.

• Decisions: I made every publish call, nine rounds of three pages, plus the names, the standards and which version of a chart was accurate. The agent proposed each one with a recommendation and the trade-off.

• Framework pages: For the library on my website, the agent wrote the pages describing the Frameworks I created, including How-to steps, FAQs, SEO fields and alt text. It built every page in WordPress and verified each one live. I reviewed.

• Charts: The agent re-rendered and rebuilt about 30 and checked the stats against sources. I uploaded them.

• Substack: The agent wrote every link line and alt text and verified every post afterward. I made the edits in 22 posts and fixed 6 links.

• Search: The agent got all 27 pages to green in Yoast, audited 74 posts and checked 31 URLs. I submitted the sitemap.

• Records: The agent kept the upload workbook, link map, tracker and status doc current. I didn’t touch them.

Look at what I actually did on that list. About half of my nine hours went to Substack edits, and another hour or two to uploading images. That’s production, not judgment. I was doing it by hand because Substack blocks the agent from that work and getting files into WordPress’s media library stayed manual.

So the division isn’t clean yet. Only two to four of my nine hours went to real judgment calls. The rest went to manual work the tools wouldn’t let the agent touch. That’s plumbing, and plumbing gets fixed. None of it needs me.

The checking was worth as much as the speed

The number that will get quoted is the hours. What the checking found mattered just as much.

Four published facts were wrong. All four were mine, and they’d been live.

• A stat credited to the wrong firm: “94% of B2B buyers use LLMs” was attributed to Forrester and CMI on two charts. It comes from the 6sense 2025 Buyer Experience Report.

• A chart that contradicted its own post: One chart said mid-market B2B was “finally off premise” in 2018. The post it ran with said the opposite.

• A chart that didn’t match its text: The post described seven questions. The chart showed six.

• A count that didn’t add up: One framework page said “four squares” and listed five.

None of these were catastrophic. Each one is exactly the kind of error a careful reader finds, and the readers of this publication are careful readers. I would not have audited 74 posts by hand. Nobody running a business has the time.

The agent found the errors. It didn’t decide what to do about them. The chart that contradicted its post needed a new label that was true in both places. The seven-question index needed its scoring range rebuilt. Those were calls, and they were mine. Verification moved to the agent. Deciding what’s true stayed with me.

Judgment becomes the lever, not the cost.

The build was fast, not frictionless, and the friction is the most useful part of the story.

The one hour I’d take back came from a call I made late. The Uphoff Value Index went from “Value” to “Valuation” and back again inside an hour. Each switch meant a chart re-render, a page retitle, a URL change and a redirect: four pieces of work, each done twice. That waste was mine, not the agent’s.

Agents multiply whatever direction you give them. Make the call early and they multiply the work. Make it late and they multiply the rework.

Not every mistake was a decision, and not all of them were the agent’s. One round of uploads landed on the development site instead of the live one. A stray bulk upload pulled in 29 files. All 18 rebuilt pages went to the trash once and came back with Undo. Every one of those was fixed in minutes, because the records were current and the standards were already written down.

That’s the operating lesson. In every transformation I was brought in to lead, execution was where things broke, and the worst breaks traced back to decisions that came after the work had started. Agents raise the stakes. They execute so fast that a clear call made up front pays off almost immediately, and a late one throws away work that’s already done. The leverage is in the calls you make before the work starts.

An Output Multiple isn’t a Scale Multiple yet

Last week I wrote about the difference between an Output Multiple, how much more a person produces on one workflow, and a Scale Multiple, how much more revenue each person brings in. This build is an Output Multiple. A large one. It is not yet a Scale Multiple, and I’d be overselling it to call it one.

For anyone new to the terms: Solo Scale is one expert running agents, from $250,000 to $5 million in revenue at 40% or better EBITDA. Team Scale is the next rung, a small senior team doing the same at $400,000 or more of revenue per employee and 25% or better EBITDA, from $2 million to $30 million. Both are defined by revenue per person, not headcount. For context, the median private SaaS company does about $175,000 of ARR per employee, and WPP, the largest agency holding company, about $130,000 per person. This build is evidence on the input side of that math, not the revenue side yet.

Then there’s the cost. The agents ran on tools I already pay for. Four weeks of the same work, hired out at typical freelance SEO rates of $75 to $100 an hour according to an Ahrefs survey, would have cost roughly $12,000 to $16,000. One thing isn’t in that comparison: my web developer made a configuration change at no charge that let the pages hold real content. It doesn’t change the order of magnitude, but it belongs in the record.

The more important point is what the comparison actually is. I was never going to hire a writer, a designer and a web producer to rebuild 27 pages of my own IP. Without the agent, the library stays half built. That’s the real Solo Scale effect. The work isn’t just cheaper. Work that would never have been funded now exists.

Whether it turns into a Scale Multiple depends on what the operator does with the hours the agent gave back. While the agent was building the library, I brought on a new advisory client. One additional client isn’t a Scale Multiple, but it’s exactly where those hours are supposed to go. If they go into more client work and better client work, the margin shows up. If they go into more production, it doesn’t.

I work more hours, not fewer

Here’s what I didn’t expect. You’d think a Solo Scale business means working fewer hours. I work more. What changed is where they go.

When I launched the firm, a real share of my week went to production: formatting, uploading, fixing links, keeping records. Most of that now sits with the agents. Those hours moved to the work clients actually hire me for: research, analysis, planning and strategy.

It shows up in the engagements. Less of my time goes to the production that comes with project work, so more of it goes to the connected issues that decide whether a business grows: positioning, packaging and pricing, the roles a team needs and who fills them, channel strategy, and coaching the leaders who have to carry it all. None of that is production. All of it sits in the Keep It box.

That’s what a services business with software margins looks like up close. It’s a judgment business with a production engine underneath it. The engine does more every quarter. The judgment is what clients hire me for.

Sort the work before you hand it off

Here is an Uphoff Advisory Framework you can download and use in your business. Put any project on it before the agent starts, and log your hours against it after.

The Judgment Lever sorts work along two lines: who executes it, the agent or the operator, and what kind of work it is, production or judgment.

• Hand it off: (agent, production). Drafting, building, rendering, record-keeping. The volume.

• Propose, don’t decide: (agent, judgment). Fact checks, audits, options with a recommendation and a trade-off. The agent brings the call. It doesn’t make it.

• Keep it: (operator, judgment). Names, standards, what’s true, what ships. Make these first.

• Fix the plumbing: (operator, production). Work you still do by hand because the tools won’t connect. This quadrant should shrink every quarter. If it doesn’t, start there.

Download The Judgment Lever →

View the full Uphoff Advisory Framework library →

What to do with it this quarter

If you’re building, running, funding or starting a services business, run one project through The Judgment Lever and keep score.

Make the “keep it” calls before production starts: names, sources, standards, what done looks like. Track where your own hours land. If most of them sit in “fix the plumbing,” you don’t have a judgment problem. You have an integration problem, and it’s the cheapest one you’ll ever solve. If you fund services businesses, ask the operators you back the same question. How much of their leaders’ time still goes to production?

The agent will take more of the work every quarter. It won’t make the calls, and it can’t make them early on your behalf. Hand it the production. Keep the judgment calls, and make them first.

Keep going: Loop Engineering: The Bridge to Workflow Transformation and The Teammate That Never Logs Off

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

Uphoff on Media is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory: a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology.

"It's not personal, it's just business" gets said most often at the exact moment it is least true. Across a table. To someone whose job, title, or plans just changed. Nobody has ever walked out of that conversation agreeing with it.

I've said some version of the opposite for close to thirty years: people who say it's not personal, it's just business are kidding themselves. Business is highly personal. And as a leader, you will need to make decisions that are in the best interest of the business and not in the best interest of someone on your team, sometimes someone you like, respect, and hired yourself.

Both halves of that are true at the same time. Most of the damage leaders do to their teams, and to themselves, comes from pretending one of them isn't.

Subscribe now

Management comes with two loyalties and no instructions

In Be Careful What You Wish For, I wrote about the question I've been asked every few months for three decades: should I get into management? That piece was about whether you're built for the seat. This one is about what happens after you say yes, when you discover the job comes with two loyalties and no instructions for holding both.

The first loyalty is to the person in front of you. The one who stayed late during the carve-out, who covered for you when your kid was sick, who you actually enjoy having a beer with.

The second loyalty is to the enterprise. That word sounds cold until you remember what it actually is: the other two hundred people, their families, the customers who bet their own plans on you delivering, the investors who funded the payroll. Deciding for the business is almost never deciding for a spreadsheet. It's deciding for a larger group of people who aren't in the room.

That's the weight. Business isn’t impersonal. It’s personal in two directions at once.

“It’s just business” and “we’re a family” fail the same way

The first lie is the one leaders tell themselves: it's just business. It's anesthetic. It lets you make the call without feeling the cost, which is exactly why it fails. The person across the table feels the cost in full, and they can tell you've numbed yourself to it. What they remember isn't the decision. It's that you seemed not to mind.

The second lie is the one leaders tell their teams: we're a family. It feels generous when you say it at the all-hands. It is a promise you can't keep. Families don't restructure. Families don't run performance reviews. Families don't sell the division. Netflix saw this clearly enough to write the opposite into its culture memo: "It's why we model ourselves on a professional sports team, not a family." Families, the memo says, "are about unconditional love." Every leader who has said "family" and then had to lay someone off has watched a normal business decision land as a personal betrayal. The word made it one.

The two lies run in opposite directions, and they fail the same way. Each one makes the tension go away by pretending one loyalty doesn’t exist. The job is to carry both.

What I told the team when we ended BYTE’s print edition

In 1998 I was group president at CMP Media, running a division north of $300 million, when CMP bought McGraw-Hill's Information Technology and Communications Group and handed the acquired brands to me. The best known of them was BYTE, which had been the reference point for serious computing since its first issue in September 1975.

From diligence onward, two groups of people were watching every move. The CMP employees in my division wanted to know what the acquisition meant for their positions and their work. The people coming over from McGraw-Hill wanted something harder to give: assurances that their brands, and their jobs, were safe.

That's the moment the family lie is most tempting. A few warm promises at an integration meeting would have bought a quiet month. They also would have been promises nobody could keep. Several of the acquired brands were in decline and hadn't led their niches for a while. BYTE's circulation had slid from roughly 523,000 in 1996 to about 443,000 by the time of the deal, the New York Times reported at the time.

We worked through the details, and then we made the calls. BYTE suspended its print edition after the July 1998 issue and moved online. A second acquired brand closed outright, its circulation and other assets folded into another title. Both decisions cost people their jobs.

When we told the team, I said what I'd believed for a long time: "'This is business, not personal' really isn't accurate. Business is highly personal, and particularly so for the people who lost their jobs today. At times, as leaders, we need to make decisions that are in the best interest of our business and that may not be in the best interests of some of the people involved. This is one of those times."

It didn't make the news any easier to hear. Nothing would have. What it did was refuse both lies in the same breath: it didn't pretend the decision was painless, and it didn't pretend the decision was still open. BYTE's last print issue still gets written about as the end of an era in technology publishing. It was. It was also the right call for the business. Both of those were true on the same day, and a leader has to be able to say both.

Managers are carrying more of this, and doing it alone

The numbers say this matters more than most leaders admit. Gallup's research has held for a decade that managers account for at least 70 percent of the variance in employee engagement across business units. As far as the team is concerned, the manager is the job.

And managers are carrying it worse. Gallup's 2026 State of the Global Workplace report puts manager engagement at 22 percent in 2025, down from 31 percent in 2022, with five of those nine points lost in a single year. The same report found leaders feel more stress, more anger, more sadness, and more loneliness day to day than individual contributors, with loneliness running ten points higher.

The person with the most influence over how everyone else experiences work is increasingly the person most alone in it. That's what happens when leaders try to carry both loyalties without a system for doing it.

Give, Ration, Hold: three zones, three different rules

The advice leaders usually get here is to be friendly but not friends, approachable but with boundaries. It's not wrong. It's just not usable, because it treats warmth and authority as one dial where turning up one turns down the other.

They're not one dial. What a leader offers the team falls into three different zones, and each one runs by its own rule.

Here is an Uphoff Advisory Framework you can download and use in your business. Use it to check which way you lean when the pressure is on.

Download Give, Ration, Hold →

View the full Uphoff Advisory Framework library →

How each zone works

Give freely. Warmth, respect, candor, your actual personality, your time when someone's world is on fire. These never get rationed, not even during a layoff, especially not then. This is the zone that keeps you yourself. The leader who goes cold to protect themselves hasn't gotten tougher. They've gotten worse at the job, because people stop telling a cold leader the truth.

Ration. Access, confidences, after-hours closeness, who hears what first, which friendships deepen and how visibly. This is where boundaries actually live. Warmth stays constant; intimacy is managed. You can be the same person with everyone and still not be equally close to everyone, and on a team, uneven closeness reads as favoritism long before anyone says so.

Hold. The decision criteria, the performance standard, the interests of the business. These get set before you need them and stay put when you do. When a hard call comes, it should read to the team as a standard being applied, not a friendship being broken. The standard does the work so the relationship doesn't have to absorb it.

The two most common failures are mirror images. The hardened leader rations what should be given, goes guarded and formal, and ends up with a team that complies without trusting. The dissolved leader gives away what should be held, bends the standard for the people they like, and ends up with a team that trusts them personally and doesn't believe any decision is fair. Most leaders lean toward one or the other under pressure. Knowing which way you lean is half the work.

The decision is business. The delivery is personal

The simplest operating rule I know: make the decision on business terms, and deliver it like a human being.

The business decision answers whether: Does this role still exist? Is this person meeting the bar? Does the plan still work? That decision should be clear-eyed, consistent, and written down before any name is attached to it.

The delivery answers how. Who tells them, and whether it’s you. How much notice. What you say, and whether you say it yourself instead of reading it off an HR script. What support comes with it: severance you fought for, a reference you’ll actually give, a call you’ll take six months later. Nothing about the delivery changes the decision. Everything about it changes what the decision means to the person living through it, and to the forty people watching how you handled it.

Two more rules sit underneath that one.

Early truth is the kind version. The cruelest thing a leader can do is protect a relationship by withholding bad news about performance. The feedback you skip in March to keep things friendly becomes the termination in October that nobody saw coming. Candor delivered early is warmth. Candor delivered late is an ambush.

Don't need your team to like you. Wanting to be liked is fine. Needing it is where every boundary breaks. A leader whose only source of support is the team will ration the wrong things and give away the wrong things, because the team has become the thing they can't afford to lose. The support system I called the hardest trait in Be Careful What You Wish For, the peers, the board member you can call, the coach, the family that knows you outside the title, is what lets you be warm with your team without needing anything back from them.

AI will make the analysis sharper. The delivery still has to be yours

There's a 2026 version of the first lie, and it's already here. Employers cited AI in 120,136 announced job cuts in the first nine months of this year, according to Challenger, Gray & Christmas, more than double the 54,836 for all of 2025. As agentic AI takes on more of the analysis behind staffing, territory, and performance decisions, leaders will be tempted to hand the moral weight to the system along with the math. The model ranked the roles. The model flagged the underperformers. It's not personal, it's just the output.

It’s the same anesthetic, and it will fail the same way. Machines can sharpen the decision. They can make the criteria more consistent and the data harder to argue with. That’s useful. They can’t carry the delivery, and a team will not accept a decision delivered by someone hiding behind a dashboard. The more of the analysis agents take on, the more of the leader's job is exactly this part: owning the call and delivering it like a human being.

What your team remembers is how you delivered it

Leadership doesn't ask you to resolve the tension between the person in front of you and the people counting on the enterprise. It asks you to carry it, openly, without numbing yourself to one side or lying to the other.

Give the warmth away every day. Ration the closeness. Hold the standard before you need it. Then, when the hard call comes, and it will, make the decision on business terms and deliver it like a human being.

It was never just business. It was always yours to decide.

Keep going: Be Careful What You Wish For and The Miss Isn't the Failure. What You Do Next Is

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don't represent the opinions of any company I've led, any board I've sat on, or any investor who's had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I'll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

"Uphoff on Media" is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory: a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Every new dollar of revenue, another card on the wall.

This fall I've heard the same story three times, from three very different businesses: a B2B agency, a PE-backed marketing services firm, and an AI consultancy. Each has rebuilt real work around agentic workflows, and the gains they describe don't round to a percentage. They round to a multiple. Ten times. Sometimes more.

The productivity is real. The payoff isn't automatic. A multiple of output is not a multiple of anything you get paid for.

If you bill by the hour and your team now does in one hour what used to take ten, you just gave your clients a 90% price cut. You worked hard for it. They didn't ask for it. And the invoice looks exactly the same as it did last year, only smaller.

Subscribe now

The number everyone quotes isn't the number that pays

Two numbers are getting blurred together in almost every conversation about agentic AI in services businesses. Pull them apart and the picture gets much clearer.

The Output Multiple: how much more a person produces on a given workflow now that agents do most of the execution. This is the multiple everyone quotes.

The Scale Multiple: how much more revenue each person brings in: revenue per employee now, divided by revenue per employee the year before agents. $150,000 a head to $300,000 is a Scale Multiple of 2. This is the number that shows up in the P&L, and eventually in the valuation.

The gap between the two is what you're handing your clients for free.

Last week, in Before You Flatten the Org Chart, I argued that revenue per employee is the number to measure before you restructure around AI. This is the next question: how did the number go up? Two businesses can post the same improvement and be building two different companies.

Revenue per employee is the right yardstick because it's the one software companies use to grade themselves, and services businesses are about to be graded the same way. The median private SaaS company now does about $175,000 of ARR per employee, according to Benchmarkit's 2026 data. The largest agency holding company runs well below that. WPP booked £11.4 billion of revenue less pass-through costs in 2024 with an average of 111,281 people, per its 2024 results: about £102,000, or roughly $130,000, per person by my math.

Margins tell the same story. The average digital agency earned a 13% net margin last year, according to Promethean Research. Its size data is the whole problem in one line: agencies under ten people run 19% margins, and agencies with more than 50 people run 8%. In the agency business, getting bigger makes the margin worse. That's been the history of the services model, where every new dollar of revenue shows up with a new salary attached.

The holding companies are raising revenue per person the old-fashioned way

The largest agencies already know the labor model is breaking, and they're responding by taking people out.

WPP went from 108,044 people at the end of 2024 to 97,388 at the end of June. In March its CFO, Joanne Wilson, told analysts the company intends to "move away from time and materials and decouple revenue from headcount." Omnicom, which bought rival holding company Interpublic (IPG), expects to end the year with roughly 105,000 people, down from about 120,000. The cuts are the core of a $1.5 billion annual savings target from the merger, $1 billion of it from labor. Its CFO added that the cuts "really have not impacted client-facing people very much at all." One holding company CEO, quoted anonymously in Forrester's Predictions 2026, put the plan in a sentence:

"By 2028, we'll double profits and halve the people."

That raises revenue per employee. It does it by shrinking the denominator, not by growing what each person produces. I've led companies through structural disruption more than once, and the pattern holds every time: cost cuts buy time, and the only thing that buys a future is a new answer to "what exactly are customers paying us for?"

Publicis is the useful exception. It grew 5.0% organically in the third quarter, after 4.8% in the second, while its peers cut. That's a data business winning share, built on Epsilon and now LiveRamp, not a smaller team doing more. Scale still works for the company that owns the data layer. It stops working for everyone billing hours on top of it.

Outside services, the gap is starker. On October 21st, Meta reported third-quarter revenue up 26% from a year earlier, with headcount down 15%, to 66,775. That’s revenue per employee up roughly half in a single year, from more revenue and fewer people at the same time. Meta isn’t an agency. It’s the comparison your board will make anyway.

The market doesn't pay for margin by subtraction

Cutting your way to a better ratio has a second problem: buyers and investors can tell the difference.

Kroll's Summer 2026 software sector update makes the point with public software companies. Median EBITDA margins barely move across growth rates: 28.9% for companies that are shrinking, 25.1% for companies growing faster than 20%. The multiples move enormously, from 2.4 times next year's revenue for the shrinkers to 7.2 times for the fastest growers. The market is paying for growth, not margin.

Services businesses won't be graded more kindly. A firm that gets to 40% EBITDA by cutting people is more profitable and no more valuable. A firm that gets there by growing revenue per person has built something a buyer will pay a multiple for.

Where your business sits on the map

Here is an Uphoff Advisory Framework you can download and use in your business. Plot your Output Multiple against your Scale Multiple and you land in one of four places.

Download The Scale Multiple Map →

View the full Uphoff Advisory Framework library →

What each square tells you

Business as usual. Neither number has moved. The agents are a pilot, a tool, or a line in the credentials deck.

Giving it away. Output is up sharply and revenue per person hasn't followed. This is where most of the productivity stories live today: real gains, priced by the hour, passed straight through to the client.

Margin by subtraction. Revenue per person is up because headcount went down, not because anyone produces more. This is the holding company playbook. It works until there's nobody left to cut.

Team Scale. Both numbers are up, and the gain shows up in the P&L. Small senior teams, agents doing the execution, revenue per employee climbing past the software median.

Team Scale is the next rung up from the model I described in Solo Scale: The New Business Model AI Just Made Possible. Solo Scale is one expert with agents, running from $250,000 to $5 million in revenue. Team Scale is a small senior team doing the same thing at roughly $400,000 of revenue per employee: $2 million to $3 million with five to eight people, up to $25 million to $30 million with 60 to 75. At the holding company rate, booking that same $30 million takes more than 200 people.

Three ways firms are getting there: build it, buy it, or sell it

The three conversations I opened with map to three different routes into that top-right box. I've changed identifying details; the math is theirs.

Build it. The PE-backed marketing services firm built a series of agentic workflows. Before, it hired three people, on average, for every incremental $750,000 of revenue. Since then it has added about $5 million of revenue, from new clients and from expansion inside existing ones, with headcount flat. Under the old ratio that $5 million would have meant 20 new hires. It made none. Its Scale Multiple will take a few quarters to read. The twenty salaries it never added show up in the P&L today.

Buy it. The agency is using an acquisition to add AI capability, with a target of taking EBITDA from 20% to 40%. That plan can go two ways, and the math shows why the route matters more than the number. Cut about a quarter of the cost base with revenue flat and 20% becomes 40%. That's margin by subtraction. Or hold the team roughly where it is, grow revenue by about 60%, and keep cost growth under a fifth, and 20% also becomes 40%, with revenue per person landing right around the Team Scale line. Same margin. Two very different businesses, and per Kroll, two very different multiples.

Sell it. The AI consultancy is building an agentic workflow app for a multi-location consumer services business with thousands of customers and a mix of full-time and part-time staff that makes scheduling and billing a manual mess. The new system takes those processes from weeks to hours. That's an Output Multiple well past ten. The question worth asking is who keeps it. The client keeps every hour it saves. The consultancy keeps only what it priced for. Bill that build as a project and the consultancy is giving it away on its client's behalf.

The Solo Scale version of "build" is my own. I built Uphoff Advisory's client tools, the Uphoff Value Index, the Sales Talent Matrix and the Marketing Talent Matrix, along with the website and the system I run engagements on, at roughly a ten-to-one ratio of output to time against what the same work would have taken before agents. I did about 90% of it myself. For the last 10%, authentication, secure data storage and wiring it all into one client portal, I brought in a developer. AI didn't replace him. It made me a much better client, because I handed him a working spec instead of a rough idea.

The judgment didn't get a discount

Scaling yourself and automating yourself are not the same thing.

Automating hands execution to something that isn't you. Scaling extends judgment that is still entirely yours to more clients, more often. Every story above is a scaling story. The research, the drafts, the versioning, the scheduling, the reporting all got cheaper. The judgment about what the work should be didn't get cheaper at all.

My Pipeline-vs-Forecast tool weights a strong verbal awaiting signature at 90%, a confirmed proposal at 50%, a contact made at 10%, and an assigned account nobody has touched at zero. No agent came up with those weights. That's decades of watching sales forecasts inflate and collapse, written down so it can run without me in the room.

Writing it down is the step most firms skip. In most agencies, taste lives in the heads of three or four people and moves by apprenticeship: a junior designer shows work to a creative director, hears "no, not like that," and slowly absorbs the standard. Agents don't apprentice. As Digiday put it, a thousand ad variants now cost what five used to. That's a production statistic. A thousand variants of mediocre is a very efficient way to dilute a brand. Taste that's codified, in standards, evaluation criteria and examples of almost-right-but-wrong, scales. Taste that lives in a creative director's head leaves when she does.

Keeping the multiple is a pricing decision

Everything in the top-right box depends on one decision: what the client pays for. As I argued in "Thanks for the Lecture on Price", most pricing gets set against cost and never gets rechecked. In a business whose cost to deliver just fell by an order of magnitude, that habit is how a productivity multiple turns into a discount.

The largest agencies see it and are moving slowly. WPP's CEO Cindy Rose said in August that outcome-based pay will "take a few years," and WPP currently has one client, Jaguar Land Rover, paying against business and brand metrics the client sets. At Pipeline360 we built the business around Demand-as-a-Service: clients paid for demand delivered, not hours worked. It works. It also teaches the lesson every services firm is about to learn: whoever defines the outcome controls the economics.

B2B marketing has already run one big experiment in outcome pricing. It was called cost-per-lead, and it taught an industry to buy the wrong outcome. Done right, it looks different: a base fee that pays for senior judgment, plus a variable piece tied to a number the client's CFO already tracks and trusts. Pipeline created. Win rate. Payback on acquisition cost. Never a metric the agency invented to make its own dashboard look good.

By 2029 the middle of the agency market is gone

By the end of 2029, the fastest-growing B2B agencies will clear $400,000 in revenue per employee. They’ll have stopped billing hours for execution. They’ll get paid on a number the client’s CFO already reports to the board.

The squeeze comes from three sides. The holding companies that own the data keep getting bigger. The AI tools keep getting cheaper. And client teams keep pulling production in-house: 82% of ANA members already ran an in-house agency as of 2023, up from 58% a decade earlier, and agents speed that up. The firm that disappears is the one in between: 50 to 500 people, billing hours for execution, already earning 8% margins, too big to be senior and too small to own the data.

If you're building, running, funding, or starting a services business, measure both numbers this quarter. Pick one workflow and calculate its Output Multiple. Then calculate your Scale Multiple for the whole business. If the first is ten and the second is barely above one, you already know where the difference went.

The multiple is already here. The only question is whose P&L it lands in.

Keep going: Before You Flatten the Org Chart, Make Sure You Can Measure What AI Delivers and Solo Scale: The New Business Model AI Just Made Possible

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don't represent the opinions of any company I've led, any board I've sat on, or any investor who's had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I'll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

"Uphoff on Media" is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory: a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

B2B buyers still build shortlists. They just don’t build them alone anymore.

Google “Tony Uphoff” and you get about 86,000 results. Thirty-five years of bylines, press releases, bios and CEO announcements make me discoverable, and by extension the two businesses I started in February. As long as someone is searching for my name.

AI is a different story. This month I asked AI the questions my audience and prospective clients ask: who to hire, who to call, what to read. 19 prompts across ChatGPT, Gemini, Perplexity and Google AI Overviews, tracked in Ahrefs Brand Radar from October 10 to 17, produced 59 answers about my two businesses and me:

• Uphoff on Media showed up in 0 of 12 category answers.

• Uphoff Advisory showed up in 1 of 43.

• “Who is Tony Uphoff?” Every engine that answered knew the career.

The models know who I am. They almost never recommend what I’ve built. That gap is where B2B buying has gone, and most companies aren’t measuring it.

Subscribe now

Search didn’t die. It changed jobs

It used to be where buyers found you. Now it’s where they check what the machine recommended. In a March 2026 G2 survey of 1,076 B2B software buyers, 51% said they now start their research in an AI chatbot more often than in Google, up from 29% a year earlier. In a Semrush survey of US B2B professionals fielded in March and April 2026, the most common workflow among those who use AI at work (41%) was to start with AI, then verify its answers through traditional search.

I’ve been brought in to transform two businesses built on the last era. At Business.com and ThomasNet.com, helping B2B buyers find suppliers was the product. The function hasn’t changed: connect the buyer and the seller. The form has. The machine does the finding now, and the shortlist comes back already built.

That’s the new sequence. The machine drafts the shortlist. The human validates it: checks the names, reads the reviews, asks a colleague, asks the model a follow-up. In the same G2 survey, 69% said a chatbot led them to choose a different vendor than they had planned, and one in three bought from a vendor they had never heard of before. The buyer still makes the call, but from the machine’s list. If you’re not on it, you’re counting on the buyer to go looking, and those numbers show how often the machine’s list wins.

For two decades the scoreboard was discoverability. SEO taught every marketing team to rank, earn the click and get found. Most companies carried that scoreboard straight into AI. They ask a model about themselves by name, get an accurate answer and conclude they’re covered. That’s discoverability in a new channel. The scoreboard that matters now is recommendation: when a buyer asks for your category, does the model put you on the shortlist?

For most companies, the pattern in my results is a false positive. A well-known founder, CEO or legacy brand makes the business feel discoverable. Ask the model about the leader and the answer sounds right. Ask it what a buyer would ask, and the company rarely makes the list. Measure only discoverability and you’ll never see the gap.

Answer engines describe the company you used to be

All three engines that described me still call me CEO of Pipeline360. I led its carve-out and built it as a standalone business, and since February I’ve served as a board advisor. Only one of the three mentioned Uphoff Advisory, the firm I run now. The sources included org-chart aggregators, executive databases and old press releases.

Answer engines default to what the web has said about your company most often, and most often usually means longest ago. If you’ve changed leadership, renamed a product, repositioned or entered a new market in the past two years, assume the models are describing your old company to your new buyers.

The stale record costs you twice. The machine uses it to decide whether to recommend you. The human uses it to check the machine’s work. Fixing that is the job of answer engine optimization (AEO), and most companies haven’t started. A Corporate Ink study reported by eMarketer in May found only 29% of marketers actively track and measure AI visibility.

The idea travels. The name stays behind

Ask the same engines “What is Demand-as-a-Service?” and all three define it accurately, each one citing Pipeline360’s own pages. I helped create it as the foundational strategy when we built Pipeline360. None of the answers mention me, and for Pipeline360 that’s exactly right: the company owns the product and the category language.

For any business that runs on a point of view, it’s a warning. The models credit whoever publishes the clearest explanation, not whoever had the idea first.

AI shortlists are built from other people’s shortlists

The one time Uphoff Advisory made a list, the question was narrow: “Who advises founders of B2B information and data businesses?” ChatGPT listed it second of four. The broad questions, the best strategic advisors for B2B media and who to call about a turnaround, returned nothing.

The citations explain why. When the models answered category questions, they rarely cited the firms or newsletters they named. They cited roundups: “best AI newsletters” and “best B2B marketing podcasts” lists from Zapier, Content Allies, Breaking B2B, The B2B Playbook and Dupple. The AI shortlist is assembled from lists other people already published.

A May 2026 study by Maryanna Franco and João da Silva, run on Friction AI’s platform, found the same mechanism at scale. They sent 14,140 queries about twelve activewear brands to five AI engines. Nike and New Balance carry the identical Google Knowledge Graph description: footwear company. Nike appeared in about 71% of athleisure recommendation answers. New Balance almost never did. In third-party content, Nike shows up alongside lululemon roughly 50 times more often than New Balance does. Equally discoverable. Only one recommended.

In The Buyer Left the Building, I argued the shortlist gets built before a human enters the conversation. In The CPM Collapse, I called 2026 the agentic inflection for B2B research and vendor discovery. This is what both look like from the vendor’s side.

Change the scoreboard first

If you’re building, running, funding, or starting a B2B information, marketing or technology business, here’s the sequence. I’m running it on my own businesses now. Move 1 changes what you measure. Move 2 serves both sides, because the machine and the human read the same record. Moves 3 through 5 earn the recommendation.

1. Change the scoreboard. Rankings, traffic and AI mentions tell you whether you’re discoverable. Add the number that tells you whether you’re recommended: how often you make the list when a buyer asks for your category. Report it first. Write 10 to 15 prompts the way your buyer asks them: the category, the problem, “who should I hire for.” Add two or three that ask about you by name. Run them across ChatGPT, Gemini, Perplexity and Google AI Overviews and track two numbers: how often you make the list, and whether the description is right. A tool like Brand Radar does this weekly. A spreadsheet and an hour a month works too.

2. Fix the record before you chase the shortlist. Read the sources the models cite when they describe you. Update the pages you control: your site, LinkedIn, leadership bios, press page. Request corrections on the ones you don’t: org-chart aggregators, executive databases, directory listings. After any leadership change, rebrand or new pricing model, assume the models haven’t heard. This is also the record your buyer reads when they check the machine’s work.

3. Publish the canonical version of your ideas. If you coin a term, a model or a method, publish the definitive explanation under your brand, with your name on it, before anyone else writes the explainer. Then make sure the people who write about it link back to yours.

4. Win the narrow question first. Pick the most specific way a buyer describes the problem you solve and own that phrase before you chase the broad category. “Advisor for founders of B2B data businesses” made the list. “Strategic advisor for B2B media” didn’t. Narrow first, then wide.

5. Get on the lists the models read. Find the 10 to 20 roundup and comparison pages the models cite in your category. The citations are visible in the answers. Earn a place on them, next to the names already being shortlisted. That’s a PR and analyst relations job. Give it an owner and a target, like any other pipeline source.

If you run a B2B media or information business, read move 5 from the other side. The lists the models read are published by someone. In the discoverability era you sold traffic. In this one, you could own the inputs to the shortlist.

If you’re funding these businesses, add one diligence question: “When a buyer asks AI for your category, how often are you on the list?” A company that’s missing there has a pipeline problem that hasn’t reached the forecast yet. If the answer comes back as a traffic chart, they’re measuring the last era.

Here is an Uphoff Advisory Framework you can download and use in your business. Run both tests on yourself and your top three competitors, then find your square. Each one tells you what to run first, what the risk is and what to watch.

Download Discoverable Is Not Recommended →

I’ll rerun the test in January

Zero and one are where my two businesses start. I know which pages to fix, which phrase to own and which lists to earn. I’ll rerun the same 19 prompts in January and publish the numbers either way.

The models are drafting your buyer’s shortlist right now, from the record the web already holds on you. Write that record on purpose, or a stale org chart will write it for you.

Keep going: B2B Branding Has a New Audience. Most Companies Are Talking to the Wrong One and B2B Thought Leadership and Content Marketing Are Broken

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, LLC, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

I’m a fan of radical thinking. It forces you to test assumptions you didn’t know you were making. You don’t always have to agree with it. Agreement was never the point.

The point is whether it makes you think. Jack Dorsey and Roelof Botha’s essay, From Hierarchy to Intelligence, did that to me. It lays out how Block is rebuilding itself as a company organized as intelligence rather than hierarchy, and it came a month after Block cut its workforce by about 40%. Most of what gets written about AI and organizations stops at copilots, a smarter assistant bolted onto the same old structure. They go further and ask whether the organization itself still needs to exist in its current form.

It pushed me past the easy cliché that AI just means fewer middle managers. Here’s where that thinking led.

The oldest constraint in business

Dorsey and Botha trace two thousand years of the same fix. The Roman legion built command around eight soldiers to a tent. The Prussian General Staff existed to route information up and down the chain. Daniel McCallum drew the first org chart in the 1850s to keep a 500-mile railroad from killing people.

Every version since, from Taylor’s pyramid to the matrix, worked around one constraint: a human brain can only hold so much context, so organizations add layers, and layers slow down information. Spotify, Zappos and Valve tried to break the pattern. All three scaled back into hierarchy or hit a ceiling.

Dorsey and Botha’s claim is that AI removes the constraint, not the org chart around it.

What Block actually built

Block is building a company world model (what the business knows about its own operations) and a customer world model (what it knows about the people transacting with it), and using both to compose solutions directly, without routing decisions through management. What’s left are three roles:

• Individual contributors who build

• DRIs who own outcomes across whatever teams a problem touches

• Player-coaches who develop people instead of running status meetings

No middle layer required.

Money is the one signal nobody fakes

Here’s the sentence that matters most in the whole piece.

“Money is the most honest signal in the world.”
Jack Dorsey and Roelof Botha, From Hierarchy to Intelligence

People lie on surveys. They abandon carts. They ignore ads.

But when they spend, that’s the truth. Block can build a company-as-intelligence because it sits on both sides of millions of transactions every day, through Square and Cash App. That gives it a signal almost nobody else has: honest, at scale, compounding.

Where this breaks for the rest of us

That’s also where the model runs into trouble for most of the people reading this. Very few B2B information, marketing, or technology companies have anything close to a money-grade signal. What they have instead is engagement data: opens, clicks, content consumption, lead scores, intent signals bought from a vendor.

I’ve run enough pipeline reviews to know how much of that gets gamed on purpose, by the same marketing teams reporting on it. The essay’s point about people lying on surveys is more true of modern B2B marketing than its authors probably intended.

A lead score isn’t a fact about someone’s business. It’s a fact about how well your content ranks.

Most B2B signal sits at the wrong end of the spectrum

That gap is the actual work before anyone in B2B tries to run Block’s playbook. Think of it as a spectrum, honest signal on one end and performative signal on the other. Self-reported data and form fills sit at the weak end: cheap to produce, cheap to fake.

Here is an Uphoff Advisory Framework you can download and use in your business. Put every number your leadership team reviews each week on it, and see how many sit right of center.

Download The Signal-Honesty Spectrum →

Engagement metrics sit a notch above: harder to fake outright, but easy to optimize for the metric instead of the buyer. Usage depth and renewal behavior sit higher still, because pretending costs the customer something real. Contract value, expansion revenue, and churn sit closest to money itself.

The businesses already living near that end of the spectrum are the ones actually positioned to build a world model worth trusting. Everyone else is redesigning an org chart around a signal that isn’t honest yet.

Everyone else is redesigning an org chart around a signal that isn’t honest yet.

This isn’t just Silicon Valley talking

It would be easy to dismiss this as some crazy idea from the tech elite. Don’t. The data says otherwise, and it’s not subtle. The number of managers and executives at public companies fell 6.1% between May 2022 and May 2025, according to Live Data Technologies.

Gartner predicted in October 2024 that through 2026, one in five organizations would use AI to flatten their structure, eliminating more than half of current middle management positions. We’re in the last quarter of that window.

I’ve been the skeptic on AI-attributed layoffs. In June I called most of it a misattribution problem: pandemic correction and PE margin engineering, relabeled as AI. That’s still true of last year’s numbers. But manager headcount specifically, not total headcount, is the number that’s moved. Nobody announced that. They just stopped backfilling the role.

This isn’t theoretical, and it isn’t just Block. GitLab and Coinbase both cut management layers this year and pointed to AI when they did. Cloudflare’s cuts reached middle management too. Google said last year it had eliminated 35% of its managers overseeing small teams. Intuit, Meta, Cisco, Atlassian, Salesforce, and Oracle all paired 2026 layoffs with AI investment or efficiency claims. Block didn’t start this. It’s just the company doing it with the strongest claim to a signal that actually justifies it.

Revenue per employee is how the market keeps score

Block’s reorg came with a scoreboard, and it isn’t the org chart. Run Block’s own numbers:

• 2025: $10.36 billion of gross profit with 10,205 people. About $1 million per person.

• The plan: $12.2 billion for 2026 with just under 6,000 people. The stock jumped more than 20% after hours.

• Now: guidance raised twice, most recently in August to $12.51 billion, after second-quarter gross profit grew 25%. About $2.1 million per person, double where it started the year.

The market wasn’t paying for the layoff. It was paying for the output per head.

Then: a vanity stat. For most of my career, revenue per employee was a number nobody built a valuation around. A 2013 analysis of public software companies found it had almost no relationship to market value. Hiring waves, sales ramp times and company age moved it more than strategy did.

Now: the cleanest read on scale. AI changed what the number measures. It shows whether technology or headcount is doing the work, and that’s the question every buyer and investor is asking. Here’s where the bar sits:

• Marketing agency average: about $163,000 per employee

• Median public software company: about $395,000

• Median AI-native software company past $100 million of ARR: $655,000 (ICONIQ, September 2026)

• Top quartile of that group: $890,000

Download The Revenue per Employee Ladder →

Why it’s hard to fake. Anyone can say AI made their team more productive. Revenue per employee shows whether the productivity reached the P&L. That’s why I added it to the Uphoff Value Index this quarter as a driver in all eight categories we benchmark. A services firm approaching the software standard earns up to a full turn of EBITDA. Two agencies with identical P&Ls, one running with 68 people and one with 40, now land nearly $1 million apart.

One caution from the operator’s chair. Revenue per employee can be gamed too. Push work to contractors or offshore partners and count only full-time staff, and the number looks better than the business is. Expect buyers to start asking for revenue per dollar of total labor cost, not just per head. Measured honestly, it’s the closest thing most B2B companies have to a money-grade signal about themselves.

Marketing and sales break first

If you’re running marketing or sales, none of this is abstract. These are the two functions most exposed, and the exposure isn’t only about headcount. It’s about who you hire, how you coach them, and what “good” looks like once the AI is closer to the honest signal than your team is.

Marketing: the funnel was never honest. Marketing has been running on performative signal for years, and the funnel proves it. Forrester has found that fewer than 1% of leads make it from the top of the funnel to a closed deal.

The industry knows the model is broken, and vendors are already pitching a replacement. Docket calls it the Agent Qualified Lead: a lead qualified by an AI agent in a live buyer conversation, capturing real intent instead of inferring it from a click or a form fill. Hiring for that looks nothing like hiring for lead-gen volume. Coaching it looks nothing like reviewing MQL counts in a Monday pipeline meeting.

The SDRs who survive aren’t dialing more. They’re coaching AI performance.

Sales: the job is splitting in two. It’s happening right underneath your comp plan. Regie.ai’s analysis of LinkedIn data found full-cycle AE seats up 10.8% while SDR headcount fell 22.7%. Two-thirds of teams cut SDRs. Fewer than one in five added them.

The SDRs who stayed are running AI on the accounts that matter, not working the phones. That’s a different hire, managed a different way, measured against a different definition of quota progress.

The playbooks most of us built our marketing and sales organizations on assumed a human was doing the qualifying, the outreach, and the judgment calls. Agentic AI is already doing large parts of all three. Keep running the old playbook and you’ll be coaching people to compete with a system that’s already faster and can’t be gamed.

The real work comes before the reorg

That’s the mistake I expect to watch a lot of companies make over the next two years: skipping straight to flattening the org before doing the harder, less glamorous work of finding or building a signal that can’t be gamed. Speed is the advantage Dorsey and Botha are chasing, and they’re right that it compounds. It compounds in the wrong direction just as fast when the thing moving fast is built on noise, not signal.

Before anyone touches their org chart, they should answer one question honestly: what’s the one number in this business that nobody can fake. If there’s a real answer, the reorg is worth having, and revenue per employee will tell you within a year whether it worked. If there isn’t, that’s the actual project, and it has nothing to do with headcount.

Keep going: AI Stopped Being a Story. Now It’s in Your Multiple and B2B Marketing Is Driving Forward, Looking Backward

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, LLC, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Somewhere in your organization right now, a team is piloting technology nobody can explain the job of. They can describe the vendor. They can describe the demo that got everyone nodding. Ask them what specific job that technology was hired to do, and watch the pause.

That pause is the whole problem. Most AI pilots that fail were never given a job. Nobody wrote down the job description the technology was hired to do, who does that work today, and what would prove it got done better. Writing it down costs nothing. Skipping it costs the pilot.

Subscribe now

Procurement gave this mistake a name. It doesn’t stay in procurement

On a recent episode of the Art of Procurement podcast, Darshan Deshmukh, President of ProcureAbility, gave this pattern its name: “death by a thousand pilots.” His diagnosis was just as plain. Organizations keep leading with the technology instead of the outcome. Philip Ideson, who co-founded Art of Procurement with Kelly Barner and built it into the source procurement leaders actually rely on, put the consequence best in his write-up of the episode: “A compelling platform or use case can be a source of inspiration, but it cannot define the destination.” (Disclosure: Art of Procurement is an Uphoff Advisory client.)

None of this is new. Procurement made this mistake with ERP, then with e-sourcing, and every other function made it with CRM. Same mistake, different software.

The ERP era supplied a half-billion-euro example. In 2018, Lidl walked away from a seven-year SAP project after spending roughly €500 million on it, according to Handelsblatt. The sticking point was not complex. Lidl valued its inventory at purchase price. The software was built around retail price. In my read, nobody settled up front which way the business actually needed that job done, and hundreds of consultants spent years trying to bend one to fit the other.

I’ve never run a procurement organization. I have spent three decades running media, marketing, and technology businesses, and every one of them had its own version of this: the CRM migration that died in month four, the martech stack nobody fully turned on, and now the AI copilot every vendor in your inbox is pitching your team. The martech version is measurable. Gartner found marketers used 58% of their martech stack’s capabilities in 2020, 42% in 2022, and 33% in 2023. The spend kept going up. The use kept going down.

The failure point is identical everywhere it shows up. The technology got approved. The job never got defined.

Why do most AI projects fail? Rarely because of the technology

The failure rate for this generation of pilots is well documented. S&P Global Market Intelligence found that 42% of companies abandoned most of their AI initiatives in 2025, up from 17% the year before, and the average company scrapped 46% of its proofs of concept before they reached production.

The more useful finding is why. When BCG surveyed 1,000 senior executives in 2024, it attributed roughly 70% of the challenges in getting value from AI to people and process, 20% to technology, and only 10% to the algorithms themselves. The part every demo is built to sell is the smallest slice of the problem.

Now look at the budget for your last pilot. If most of it went to the vendor, you funded the 30% and left the 70% to chance.

McKinsey’s August 2026 State of AI survey points the same direction from the other end. The companies getting real earnings impact from AI are a small group, about 6% of respondents. Nearly three-quarters of them say they have fundamentally redesigned workflows around it. Among everyone else, the figure is about a quarter. The winners started with the work. Everyone else started with the tool.

The business case already exists. It was built backwards

Almost every one of those pilots cleared some kind of approval process. Somebody wrote a business case. There were ROI projections, a vendor comparison matrix, maybe a proof of concept demo that went well. This is not usually an absence-of-diligence problem. It’s a diligence-aimed-at-the-wrong-target problem.

The business case documented what the technology could do. It rarely documented, in plain words a non-technical person in the business could repeat back accurately, the job the organization actually needed done. Capability and job are different questions. Most business cases answer the first and assume the second.

Pilots multiply for another reason nobody puts in the business case. A pilot is the one way to say yes to the vendor and no to accountability in the same meeting. Nobody’s name is on it. Nobody set a date to decide. It costs less than the argument about killing it would.

Death by a thousand pilots is death by a thousand decisions nobody made.

A job is not a use case

I’ve made a version of this argument before, about customers rather than internal technology decisions. When I was brought in to transform ThomasNet.com, my team thought we understood the dozen or two things our customers were trying to accomplish on the platform. The real number, once we did the research, was over 350. We had been building for what we assumed the job was, not what it actually was. I wrote about what that discovery did to the company, and to my own thinking, in the post You Don’t Know What Your Customers Are Actually Trying to Do.

The same blind spot shows up on the buy side, and it may be more dangerous there, because nobody outside your own walls is going to do that research for you. A vendor’s use case library describes what other companies did with the tool. It does not describe the job inside your organization, under your specific circumstances, that this technology is being asked to do. If you cannot state that job without naming the technology, you haven’t defined the job. You have named a category.

A better demo makes the mistake easier to make

An ERP demo in 2004 was a slide deck and a promise. An AI demo in 2026 is the product working on your own data, in your own language, in front of the people who will approve it. The gap between the demo and the job has never looked smaller. It has rarely been larger.

A demo shows one job done well, under conditions the vendor chose. Your organization runs hundreds of jobs, under conditions nobody chose. That was the lesson of the 350 at ThomasNet, and no demo is built to show it to you. A good demo makes defining the job feel unnecessary. That is what it was built to do.

So I have a rule, and vendors do not love it: no demo until the job is written down.

Two companies wrote two different sentences

Klarna is the clearest recent example of what happens when the job gets written wrong. In early 2024 the company said its AI assistant was doing the work of 700 customer service agents and handling two-thirds of its customer chats. By May 2025, CEO Sebastian Siemiatkowski told Bloomberg the company was investing in human support again. His explanation was unusually candid: “As cost unfortunately seems to have been a too predominant evaluation factor when organizing this, what you end up having is lower quality.” The job got written as “cut the cost of support.” The job customers needed done was “get my problem resolved.” The pilot did the first job and failed the second.

Morgan Stanley did the opposite. Before rolling out its GPT-4 assistant to wealth management, the firm started with a narrow job: help a financial advisor find the right answer, fast, in a library of roughly 100,000 research documents. Every use case was tested against real advisor tasks, with advisors grading the output, before it shipped. According to OpenAI’s case study, more than 98% of advisor teams now use it. Same class of technology as the pilots in the S&P numbers. The difference was a job defined before the purchase order was sent.

I wrote the jobs down before I hired a developer

When I set out to build the decision tools I use in my advisory work, the easy path was obvious: pick a fast AI development platform, sketch a rough concept, and let a good demo carry the rest. I did the opposite. Before I hired a developer, I wrote out the job to be done for each tool.

The valuation tool’s job: when a CEO is weighing a major decision, show what that decision does to the value of the business, and on what timeline. Not “calculate what the company is worth.” That distinction decided what got built. A valuation calculator answers a question once. A decision tool gets checked before every big call, so it needed drivers an operator can actually move, benchmarked against real trading multiples.

The talent tools’ job: when a manager has to hire, coach, or restructure a sales or marketing team, help them make that call on evidence, starting with telling a strong resume from a strong hire. Writing that down ruled out the obvious product, an employee-tracking system. The job is the decision, not the monitoring.

The technology, the specific platform, the build sequence, all came after those jobs were written down, not before. That ordering is the entire difference between a tool people keep opening and one that stops getting used after the second demo.

Agentic AI removed the friction that used to slow this mistake down

The old version of this mistake had natural speed bumps built in: long procurement cycles, an IT approval queue, a budget committee that met quarterly. A bad idea had time to die on its own before real money moved.

Agentic AI tools mostly skip all of that. A capable pilot can now show up as a subscription one manager puts on a corporate card, no committee required. Zylo’s 2026 SaaS Management Index found expensed software spend up 267% in a year, with ChatGPT now the single most expensed application, and business units, not IT, controlling 81% of software spend. Gartner expects more than 40% of agentic AI projects to be canceled by the end of 2027, and its analysts describe most current projects as experiments “mostly driven by hype.”

The old failure mode took a fiscal year to play out. This one takes a sprint. A twenty-year-old mistake now compounds at the speed of a subscription click, and most operators have not updated their discipline to match.

The instinct is to add process: a committee, a review, a form. That loses. Governance that costs more than the subscription it governs gets routed around, by the same manager, with the same card. Whatever discipline you add has to be cheaper and faster than the pilot it gates. That is why the test below is three questions and not a forty-page deck.

Three gates, cleared before the PO gets signed

Here is the test I now run on every technology pitch that reaches my desk, whether I am advising a client or approving something inside my own business. I call it the Application Test: three gates, cleared in order, before a dollar moves.

1. State the job. In plain words a non-technical person in the business could repeat back accurately, with no vendor name and no product name in it.

2. Name the owner and the baseline. Who is doing that job today, and how well, in a number you already have: time, cost, error rate, or volume. A committee is not an owner.

3. Set the 90-day signal. The one measure you will check in ninety days that tells you the job is closer to done, separate from whether the pilot is still running. Logins, seats, and prompts measure the tool, not the job. Decide in advance what you will do with the answer: scale it, fix it, or stop it.

Fail any gate, and you do not have a business case. You have a good demo and an open checkbook. Klarna’s pilot would have cleared gate one on cost and failed gate three the first time someone checked whether customers were getting their problems solved.

The Application Test decides whether a pilot earns its first dollar. Whether it earns the second is a different discipline, the one I laid out in Your Agentic AI Isn’t Failing. Your Math Is.: never fund the next stage without naming what it has to pay back.

Here is an Uphoff Advisory Framework you can download and use in your business. The full worksheet adds the checkboxes, the red flags, and the scoring. Print it, hand it to the team bringing you the next pitch, and ask them to fill it in before the demo, not after.

Download The Application Test →

BCG put the algorithms at a tenth of the problem and people and process at 70%. In my experience, most of that 70% starts with a job nobody wrote down.

Here is the fastest way to find out whether that is true in your business. Pull up the last three pilots you approved. Write down the job for each one now, after the fact, with no vendor name in it. If you can do it in under a minute, the pilot probably has a future. If you cannot, you already know how it ends. You just haven’t admitted it yet.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Paramount Skydance closed its acquisition of Warner Bros. Discovery on October 6. The combined company is now simply called Skydance: roughly $70 billion in revenue, more than $30 billion a year in content spend, and about $79 billion in debt.

Most of the coverage will treat this as a Hollywood story. Studio lots, franchise slates, who runs HBO. That misses the point. This is a technology bet with a studio attached, and David Ellison has made that his whole argument:

“Unless you can build a tech product truly competitive with Silicon Valley, you can’t compete.”

You might be asking how the merger of Warner Bros. and Paramount is a B2B story. It's the clearest case study I've seen this year of a legacy business adapting to technology disruption. The real test isn't the deal. It's whether the company can change how it operates, and the order it has always done things in.

Ellison is right about the problem. Whether he can fix it comes down to a sequence the traditional studios have gotten backwards for most of the streaming era.

Subscribe now

Netflix won on the how, then the who

Netflix didn’t take over filmed entertainment because it had better content. It didn’t produce award-winning content until 2013, when House of Cards won its first Emmy. Netflix was 16 years old. It won because it started with the how: consumer choice, ease of access, and a user experience built by technologists, not bolted on by a distribution department.

Only after it owned the how did Netflix go after the who: world-class creators on screen and behind the camera. By then it had something no studio could offer talent. It knew exactly who was watching, what they finished, and what they’d watch next. The how made Netflix the most attractive place for the who to work.

That order matters. Netflix still doesn’t own a cloud. It rents Amazon’s, and still spent $3.4 billion on technology and development last year, because the advantage was never the servers. It was the product organization that decided what got built.

The studios bought the business model but kept their old operating model

The traditional studios watched this happen and responded by launching streaming services. Subscription pricing, original programming, a global app. They copied the business model.

They never changed how they worked. The operating model stayed talent first and talent last, with user experience treated as somebody else’s job. The apps launched feeling like 2009 enterprise software draped over a content library. Warner Bros. Discovery renamed HBO Max to “Max” in 2023, erasing one of the most valuable brand names in media, then put it back two years later. That isn’t a technology company making a product decision. That’s a studio treating the one of the most valuable consumer brands as an afterthought.

Every prior owner of Warner Bros. made a version of the same mistake. I traced that history in Bundling Didn’t Die. It Moved.: AOL, AT&T and Discovery each arrived with a distribution or cost thesis, and none of them changed how the place actually worked.

Skydance is the first studio organized how-first

Look at what Ellison built before the deal closed, and almost none of it is about movies.

Paramount+, Pluto TV and BET+ are moving to a common Oracle backend. Skydance Animation already runs Oracle’s “Studio in the Cloud” production model, and Ellison intends to scale it across every production workflow. Safra Catz, Oracle’s former CEO, joined the board. Dane Glasgow, a former Meta and Google product executive, now runs product, digital platforms and AI. Larry Ellison personally guaranteed roughly $45.7 billion of the equity.

The closing announcement reads the same way. More than $6 billion in run-rate synergies over three years, with technology the first source named. HBO Max and Paramount+ unified “into a single service over time.”

And the org chart. Ynon Kreiz, the former Mattel CEO, became co-CEO on October 5 and owns day-to-day operations and integration. Ellison kept strategy, creative, talent, technology and capital allocation. Other than Netflix, no studio has made technology the centerpiece of its strategy the way Skydance is.

Cloud + AI + IP is the how. It still needs the who.

Ellison’s formula is Cloud + AI + IP. The first two terms are the how. The third is what the who produces. And the plus signs hide the real math: this works as multiplication, where a zero anywhere zeroes everything.

Cloud is the easiest term to buy. Moving to Oracle can lower the cost to serve. On its own it builds nothing a competitor can’t rent next quarter. One catch: Oracle makes money when Skydance consumes more cloud, so watch whether that relationship gets priced and measured like any other vendor’s.

AI is only as good as the data under it. Skydance today has two flagship streamers on two technology stacks, a free ad-supported service, a stack of cable networks and two newsrooms. That’s a lot of customer data and no single view of the customer. Netflix’s how was built on one identity, one product, one recommendation layer. Until Skydance has that, AI gets applied to fragments.

The who is the term studios think they already have. They don’t, not anymore. HBO’s value was built on creators believing they’d be protected, and the Zaslav years eroded a lot of that trust. That’s why the leadership moves on the eve of closing matter. Casey Bloys, the steward of the most trusted creative brand in the building, is set to run all of streaming. And Kreiz is the executive who turned Mattel from a company that sold plastic into one that monetizes characters. The Barbie movie grossed more than $1.4 billion worldwide. Mattel's direct take was about $125 million, and Barbie toy sales rose 27% in the fourth quarter but just 3% for the year. The ticket sales were never the point. The movie proved Mattel's characters could carry a franchise, and that's the asset Kreiz knows how to build.

Here’s the tension Ellison has to manage. The how-first instinct, applied by a company carrying north of 6x EBITDA in debt, points straight at cost. Fitch cut Paramount to junk in March, and the stated target is 3.0x leverage by the end of 2029. Every quarter between now and then will reward using technology to cut production cost. Netflix used technology to make the experience better, which is what made it the place creators wanted to be. Same tools. Opposite outcomes.

The line between a technology company and a tech-enabled one is disappearing

Netflix disrupted Hollywood as a technology company. It built its own product, its own recommendation engine and its own engineering culture, and it got valued like a tech company. Content was what the platform needed.

Skydance is something else: a tech-enabled studio. It owns the IP, rents the infrastructure from Oracle, and hired product leadership instead of building an engineering organization from scratch. A few years ago, that structure all but guaranteed second-place software. Agentic AI changes the math. When agents write the code, run personalization and handle the workflow, capabilities that once took a large engineering organization can be rented and managed by a much smaller team.

That’s the real experiment inside this deal. Skydance doesn’t need to become Netflix. It needs Netflix’s results from a different structure: IP it owns, technology it rents, and agents doing work that used to take an engineering department.

Boards and investors are asking every B2B information CEO a version of the same question: do we need to become a software company? In the agentic era, the answer is increasingly no. You need to operate how-first. Who builds the stack matters less every year. Who decides what it’s for matters more.

The same sequence error runs through B2B

I’ve covered Oracle as an analyst for Cloud Wars, and I’ve spent most of my career running businesses that owned valuable content and rented the technology. B2B made the studios’ mistake first.

B2B information companies digitized the business model years ago. Print subscriptions became digital subscriptions, then SaaS pricing, then data licenses. Most of them never changed how they worked. Editorial first, sales first, product experience somebody else’s problem. Plenty of B2B companies still have world-class experts and analysts sitting behind a product their customers tolerate rather than use.

When we rebuilt ThomasNet from a 120-year-old industrial directory into a data platform, the infrastructure decision was the easiest one we made, and in the end, the step that opened the door to growth and value. We moved legacy, on-premise systems to the cloud and went all in on AWS, the same cloud Netflix rents. The hard part was changing the focus of the business: from obsessing about a declining base of advertisers to building the product and user experience around how buyers actually searched and sourced. Then we put a century of supplier content to work inside that experience.

Agentic AI is about to make this sequence error expensive. When the buyer is an agent, the how is the whole front door. If the machine doing the buying can't find, read and use your content, it never gets considered. The who doesn't matter if the how can't get it through the door.

Here is an Uphoff Advisory Framework you can download and use as a diagnostic tool in your business. Plot where you are on the how and the who today, and which one you’re actually investing in.

Download The How/Who Matrix →

Four signals will show which sequence Skydance is running

1. Does the Oracle migration show up as a better product or only a cheaper one? Cost to serve is half the story. Watch the app.

2. When does one subscriber identity span HBO Max and Paramount+? “Over time” isn’t a date.

3. Do the first AI announcements cut production cost, or create new value from existing characters and catalog?

4. Does HBO’s talent stay? The creative deals signed over the next 18 months will answer that before any press release does.

Ellison has the capital, the cloud relationship and a credible division of labor at the top. That’s more than any prior owner of Warner Bros. brought. But the studios already proved you can adopt the business model without changing the operating model. Skydance’s bet is that a studio can learn Netflix’s order of operations: the how first, and the how in service of the who. If it does, this is the first legacy media deal that runs like a technology company without becoming one. If it doesn’t, it’s the most expensive cloud migration in history.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

The Uphoff Value Index tracks what buyers pay for B2B information, marketing and technology businesses, and why it changes. Each quarter, this report shows what moved and what it means if you're building, running or funding one.

What the UVI is, and who it’s for

What it is. The Uphoff Value Index is a valuation benchmark for B2B information, marketing and technology businesses. It starts where a buyer starts, with an EBITDA multiple range for each of eight categories, then estimates where a company sits in that range based on three factors:

• Growth and margin, read together, so strength in one can offset the other

• Revenue per employee, measured against the category’s own peers

• Protection against AI disruption, new this quarter

The eight categories: B2B Media · Media & Events · Business Intelligence · Data · Marketing Services · MarTech and AdTech · Demand Generation and Content Syndication · B2B Events

What it’s built on. Three sources: disclosed M&A transactions; the trading multiples of a fixed set of public companies in each category; and research from the advisers who run these sale processes. This quarter that includes Collingwood’s 2026-27 Market Report, a survey of about 100 active acquirers across information, media and events, and the source of the deal counts and buyer findings below.

The ranges are checked every quarter and fully reset at least once a year. A range moves only when the evidence moves it by half a turn of EBITDA or more, so a company’s number changes when the market does, not when a stock has a bad week. The UVI Review Panel, a group of operators, investors, investment bankers and industry analysts who work across these categories, reviews every update before it ships.

How you use it. The Uphoff Advisory client tool pairs that estimate with a diagnostic of the five Value Drivers that build what a business is worth over time:

Revenue Quality · Market Position · Operational Leverage · Leadership Depth · Exit Readiness.

This quarter’s ranges, the new AI protection score and revenue per employee for every category go live in the client tool on uphoffadvisory.com later this month.

Who it’s for. Operators running these businesses, investors funding them, and founders building them. As I argued in How Much Is My Business Worth Is the Wrong Question, the exit multiple is a report card on decisions made years earlier. I’ve been CEO of five businesses, each one a turnaround or a transformation. Every time, the value the next owner paid for was built or spent long before anyone talked about a sale. Most leadership teams learn what their business is worth when a buyer tells them. I built the UVI to show them while there’s still time to change it, and to line up today’s decisions with the value they want to create and the exit they’re planning for.

Subscribe now

Q3 2026: six of eight categories repriced

The dividing line is AI exposure at the revenue line. Data, B2B Media and MarTech fell because AI is already visible in their revenue. Business Intelligence held its ceiling because buyers keep paying for the data AI needs.

The deal counts agree. Information deals rose 88% in the first half of 2026, from 51 to 96, and events deals rose 31%. Media deals fell 46%, from 54 to 29.

Size now shows up in price. The year’s headline deals carry a scale premium most businesses in these categories won’t see, so two ranges moved down to price the lower middle market.

AI defensibility has become the gate. Buyers test it before they look at growth or margin. On average, 60% of the sale processes buyers entered never closed, and more of them now end quietly before launch.

The UVI Index, Q3 2026

Where AI reached the revenue line, the multiple followed

Data. In July the debate about contact and intent data was whether the market had priced AI disintermediation too early. ZoomInfo's second quarter settled it: revenue up 1.2%, a full-year outlook that still points to a decline of about 3%, a $650 million impairment, and the stock down about two-thirds in a year. Buyers have stopped discounting a future threat and started discounting revenue that no longer grows. ZoomInfo’s 35% adjusted operating margin doesn’t earn a premium on a flat top line. I looked at what that means for go-to-market data businesses in Monday’s post, B2B Marketing Is Driving Forward, Looking Backward.

B2B Media. The category's best public reference point used to be TechTarget, with margins north of 30% before the Informa combination. Informa TechTarget now runs 10% to 13% on revenue down 3%, and it has written off nearly $1 billion of goodwill since the start of 2025, including $45 million in the first quarter of this year. Buyers value digital media businesses with $5 to $10 million of EBITDA at 6.5 to 9.5 times EBITDA. Advertising-led media lost its best argument for a higher multiple. The value hasn't left the category. It's moving to the businesses that turn an audience into data a buyer can use. Media companies with real data revenue already price that way, and they belong in Business Intelligence.

MarTech and AdTech. The ceiling moved because the company holding it up moved. The Trade Desk set the top of this range for years and now trades around six times EBITDA, with the stock down three-quarters in a year. Zeta Global, growing 44%, trades around 19 times forward EBITDA, but a single company growing at that rate is an outlier. Criteo’s falling revenue marks the floor. In this category growth is the multiple, and very few businesses have it.

Intelligence held its ceiling. In events, scale is what gets paid.

Business Intelligence. Two strategic buyers paid up for data that AI agents need: Publicis for LiveRamp at about 12.3 times forward adjusted EBITDA, and Nielsen for DoubleVerify at about $2.15 billion. DoubleVerify’s growth had slowed to under 3%, so the low end of the range widened. Proprietary data still commands a premium. Slow growth still costs you, even when the data is good.

Events. Emerald, CloserStill and Hyve priced at an estimated 11 to 14.5 times EBITDA, and the same handful of private equity firms is now buying events businesses from each other. That's what scarcity looks like at the top of a market. Most events businesses aren't at the top. Buyers value organizers with $5 to $10 million of EBITDA at 9 to 12.5 times for tradeshows and 6 to 9.5 times for conferences, and the largest buyers are already refusing to look at anything under $2 or $3 million of EBITDA unless it fits closely with what they own. The demand for events is real. An agent can't replicate a room full of decision-makers. The UVI range moved down, from 10 to 16 times to 8 to 14, to price the business most operators actually run.

In services, the spread inside the range is the story

Demand Generation and Content Syndication. No new disclosed transaction this quarter. Public comps and private-market buyer conversations point the same way: demand generation and sponsorship businesses owned by media companies trade at 6.5 to 10x. Two things hold that pricing down: the hangover from the pandemic boom, when event budgets moved to leads and then moved back, and a flood of low-cost lead providers that has taught buyers to discount lead quality across the category. The top of the 7 to 12x range comes from the exception, and the exception is real. Intent-led businesses with recurring revenue that can show which accounts are in market and prove the return are starting to pull away from media pricing. My read: the market is two to three years from treating that revenue as subscription-like, and pricing it that way. (Disclosure: I'm a board advisor to Pipeline360, which competes in this category.)

Marketing Services. The range held at 8 to 14x. What changed is how buyers place a firm inside it. The holding companies set the ceiling for generalists: Omnicom trades at about 7.8x EBITDA, which limits what the biggest buyers can pay for an agency that looks like what they already own. The premium goes to firms that don’t look like that. The average agency does about $163,000 of revenue per employee. The median public software company does about $395,000. Agentic workflows are closing that gap one firm at a time, and buyers pay for the firms that close it.

Take two agencies with identical P&Ls: $11 million of revenue, $2 million of EBITDA, 15% growth. One runs with 68 people, the other with 40. Last quarter the index valued both at 9.8x. Now the 40-person firm is worth 10.3x, about $1 million more on the same earnings. Services is where the gap is widest, and where it will reprice firms fastest.

Revenue per employee now moves the multiple

Revenue per employee is the cleanest single read on whether technology or headcount is doing the work. Starting this quarter, the UVI measures it in all eight categories, each against its own public peers, because the bar isn't the same everywhere. Listed B2B media companies run about $260,000 to $370,000 per employee, research and advisory firms about $280,000 to $320,000, and MarTech and AdTech platforms about $390,000 to $780,000. A firm at its category's baseline gets no adjustment. A firm approaching the top of its peer range earns up to a full turn of EBITDA, and one well below takes a discount. A B2B media business doing $333,000 per employee picks up more than half a turn, about $1.7 million on $3 million of EBITDA.

Growth and margin set the multiple. AI defensibility decides whether you get one.

For years, growth plus margin explained most of the difference in what these businesses were worth. It explains less now. Companies with similar growth and similar margins are trading far apart, and the gap comes down to how defensible they are against AI. In Collingwood’s survey of about 100 buyers, 91% called the AI resilience of the business model essential or important. 86% said the same of the end market, and 74% of a direct relationship with the customer. Buyers now sound out businesses six to twelve months before a sale, and 60% of processes never reach a close.

So starting with this quarter’s update, the UVI scores AI defensibility. AI Disruption Protection asks four questions of every business in every category:

1. Substitution. Could AI reproduce what customers pay you for?

2. Customer exposure. How exposed is your customers’ own business to AI?

3. Relationship ownership. Do you own the customer relationship, or rent it from search and platforms?

4. Data moat. Is your data hard to recreate and built into how customers work?

The adjustment is deliberately uneven. A well-protected business earns up to a full turn of EBITDA. An exposed one loses up to a turn and a half, because buyers treat AI defensibility as a gate, not a sliding scale. Two data businesses with identical P&Ls show the effect: $5 million of EBITDA, 5% growth, a 20% margin. The one buyers believe is protected is worth 5.6x, or $28 million. The one they think AI can route around is worth 4.0x, or $20 million, and it may never get a process started.

What this means if you’re running, funding or building one

If you’re running one: Find your category and be honest about where you sit in it. The market stopped paying for margin without growth this quarter. If your margin is strong and your growth is flat, that’s no longer a premium story. Fix the growth before you plan a process. Work through the four AI questions before anyone calls, because buyers form their view on AI early and most processes never close. Then track revenue per employee and cash conversion with the same discipline as EBITDA. Buyers already are.

If you’re funding one: The discounts in data and media aren’t automatically opportunities. Some of these businesses are cheap for a real reason. Underwrite the revenue line and the cash it actually produces. Ask for revenue per employee before you ask for EBITDA, in every category, and benchmark it against the right peers. And watch the gap between public and private prices: listed information companies are down 20% to 30% from their late-2025 peak, private values for the best assets have held, and historically the private market follows with a lag.

If you’re building one: Build where the multiple holds: proprietary data an agent needs and can’t scrape, experiences an agent can’t replicate, or a services model where each person carries software-level revenue. Own the customer relationship from day one, because renting it from search and platforms is now a discount. And build toward the size buyers will look at. Below $2 to $3 million of EBITDA, your value depends on fitting a specific buyer, so know who that buyer is.

The market isn’t guessing about AI anymore. It’s reading the revenue line. Make sure yours says what you think it does.

Subscribe now

Sources and methodology

How the ranges are set. Each category’s EBITDA multiple range is built from three inputs read together: disclosed M&A transactions, trading multiples for a fixed set of public companies in each category, and buyer research from the advisers who run these processes. Public multiples come from a single source, Koyfin (S&P Capital IQ data), using trailing twelve-month EV/EBITDA on reported earnings, read October 1, 2026. Ranges are checked every quarter and fully reset each spring. A range moves only when the quarter-over-quarter change in the evidence reaches half a turn of EBITDA or more.

How a company is placed in its range. Growth and margin are combined as a single Rule of 40 score, so strength in one can offset the other. Revenue per employee adjusts the result against the category’s own public peers. An AI disruption protection score, built from four questions on business model, end market, customer relationship and data, adds a capped premium or discount. Every adjustment is capped, so no single factor can override the evidence.

Revenue per employee peers. Trailing twelve-month revenue divided by reported headcount, from company filings, October 2026. Figures are shown in the revenue per employee chart in the Marketing Services section.

Deals and company data cited

• ZoomInfo Q2 2026 results and impairment

• Informa TechTarget Q2 2026 results

• Zeta Global Q2 2026 results

• Criteo Q2 2026 results and guidance

• Publicis acquisition of LiveRamp

• Nielsen acquisition of DoubleVerify

• Hellman & Friedman acquisition of Hyve

• CloserStill sale

• Informa TechTarget Q1 2026 results ($45.0 million goodwill write-down)

• Informa TechTarget Q2 2025 results ($841.3 million goodwill write-down, first half of 2025)

Research and benchmarks

• Collingwood, Market Report 2026-27 (survey of about 100 active acquirers). Coverage in A Media Operator

• Marketing agency revenue per employee (Promethean Research, via Haus Advisors)

• Public SaaS revenue per employee (Benchmarkit 2026, via The SaaS CFO)

Figures are rounded and current as of October 1, 2026. Examples in this issue were run through the UVI client tool. Nothing here is investment advice.

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Ninety-four percent of B2B buyers now use AI somewhere in the buying process. Forrester and 6sense surveyed buyers separately and landed on the exact same number. Eighty-nine percent of B2B marketers use AI to create content

Same technology. Opposite jobs.

Buyers use it to decide. It does the research, narrows the field, and builds the shortlist, and 95% of purchases still come from that initial shortlist. Marketers mostly use it to produce: more content, more sequences, more campaigns, aimed at a buyer whose AI is now filtering most of it out.

That’s an industry driving forward while looking in the rearview mirror. And the market has started pricing the difference.

Subscribe now

The questions were about leadership. Now they’re about value.

In January I published 3 Questions B2B Marketing Still Hasn’t Answered:

1. Ownership. Who actually owns demand creation and is accountable for revenue?

2. Demand quality. What does quality mean when buyers are anonymous and AI-driven?

3. Outcomes. Can the industry move from fragmented marketing tools and systems to trusted, outcome-based systems?

It’s still a leadership problem. Nine months later it’s a valuation problem as well.

ZoomInfo took a $650.5 million goodwill impairment in its second quarter, on revenue growth of 1.2%, and guided to a full-year revenue decline. In September, Clay, which sells signal and workflow to the same go-to-market teams, raised money at $7.1 billion, more than double the valuation it carried thirteen months earlier, after quadrupling revenue in 2025. Same buyers. Same category. Opposite valuations.

It shows up inside single companies too. Salesforce stock was down 26% by early February, as software shed roughly $1 trillion of market value. By late August, its Agentforce and Data 360 recurring revenue had reached nearly $3.9 billion, up more than 210% in a year. One P&L, moving in two directions at once.

Agentic AI doesn't lift or sink a sector. It splits product lines on two variables: what you get paid for, and whose AI you're building for. Most of the industry is getting the second one wrong

Question #1: The buyer's AI now owns the front door of demand creation.

The January question asked whether marketing, sales, or ops owns demand creation. The answer came from a direction none of them were watching. The buyer owns it, and the buyer's machine does more of the work every quarter.

Here is the counter-intuitive part. Buyers are reaching out to sellers earlier than they used to: about 61% of the way through their journey, down from 69% a year before, according to 6sense.

They still start that conversation themselves roughly 80% of the time. And 95% of purchases still come from the shortlist they built before the first call. Earlier contact hasn’t reopened the field. Buyers are calling to check a decision they’ve mostly made.

Gartner expects AI agents to intermediate more than $15 trillion of B2B spending by 2028. Even if Gartner is off by a year, the direction isn't in question. So the ownership question has changed shape. The issue now is who owns what the buyer's AI knows, believes, and repeats about you before the buyer ever raises a hand.

In most marketing organizations, nobody does. A Scribewise and Scrunch survey of 602 marketers this spring found 91% claim a documented AI search strategy, while 62% are still measured on traditional SEO metrics. The job moved. The scoreboard didn’t.

Question #2: Quality is whatever the buyer’s AI trusts.

In January I wrote that the industry still sells leads while buyers behave like a buying group. The buying group just picked up a member nobody counts: the agent doing the research. Forrester's 2026 research puts the typical decision at 13 internal stakeholders and nine external influencers. When the product itself has generative AI features, the group doubles, from seven people to 14. A lead is one person. An agent doesn't fill out a form.

For twenty years, the seller defined quality: a scoring model, an MQL threshold, a handoff rule. The buyer's AI defines it now, and it isn't grading on machine logic alone. It rewards what's verified across independent sources, specific enough to answer a real question, and solid enough that someone in the buying group will stand behind it. It discounts volume.

Volume is where most seller-side AI went. A 10Fold survey of 400 B2B tech marketers found 46% producing three to five times more content than the year before, on budgets up 10% or less. The Content Marketing Institute found 87% of marketers using AI for content report productivity gains, and 39% report better content performance. Output multiplied. Results didn’t.

Compare what happens when quality measurement follows the buyer. Palo Alto Networks dropped MQLs for buying groups and, per a Forrester case study, saw a 17% higher win rate, with multi-stakeholder opportunities eight times more likely to advance.

The value math follows. Anything sold by a unit the buyer’s AI makes redundant (contact records, gated-content leads, undifferentiated intent, outbound volume) is being repriced down. Verified signal that a buyer’s agent needs and can’t scrape is being repriced up. The market is paying for signal and marking down records.

Question #3: Trust moved from the tool to the result.

The third question asked whether the industry could move from fragmented tools to outcome-based systems. The software market answered for it this year. As I argued in The Stack That Broke Marketing, the fix for sprawl was never more technology. It was a different application of it.

Buyers have stopped trusting claims and started testing them. More than 60% now run a trial before they buy, according to Forrester, and only about a third of those end up buying from the vendor whose trial they ran. The trial has become the place where shortlists flip.

And buyers trust the machine about as much as they trust you. Gartner found 51% of buyers worry about misleading information from generative AI, and 49% worry about misleading information from sales reps. So they check one against the other: 69% want a sales rep to validate what the AI told them.

Trust sits with whoever is accountable for the result and can prove it. Features, attribution models, and dashboards still don’t answer the CEO’s question from January. Evidence does.

Both org charts are being redrawn. Only one side is doing it on purpose.

The buyer’s side added people to check the machine. Buying groups double when AI is part of the purchase. Procurement is a decision-maker in 53% of buying cycles, per Forrester, and is increasingly involved from the earliest stages. 6sense found 58% of buyers engaged sellers earlier specifically to evaluate the vendor’s AI. Buyers built a verification layer, and it is getting larger.

The seller’s side added machines and subtracted titles. Spencer Stuart found that 31% of S&P 500 companies no longer have an enterprise CMO, and the ones who remain average 4.1 years in the job. Postings for GTM engineers, the people who wire data, AI, and workflow into the revenue engine, grew 205% year over year to roughly 3,000 open roles in early 2026, according to one job-market analysis, and 91% of the companies that have hired one employ exactly one. Gartner expects AI agents to outnumber sellers ten to one by 2028, while fewer than 40% of sellers will say the agents made them more productive.

The money follows the same pattern. Gartner’s 2026 CMO Spend Survey puts marketing budgets at 7.8% of revenue, with 15.3% of that going to AI. Only 30% of those CMOs say their organizations are ready to use it, and 56% say they lack the budget for their own AI strategy.

Both sides are running into the same wall. Gartner’s procurement research calls it a productivity paradox: AI makes individuals faster, and the gain shrinks at the team level. The buyer’s side is answering with more judgment. The seller’s side is answering with more output.

One side is building a windshield. The other is polishing the mirror.

Five predictions

Here is what I expect to happen from here, and what I’m watching for. Each one cuts against the consensus that AI makes B2B simpler, cheaper, and less human.

1. AI is making B2B buying more human, not less. Groups double when AI is involved. Two-thirds of buyers say they prefer a rep-free experience, and more than two-thirds want a rep to check what the AI told them. Both are true at once. The seller’s job is moving from persuasion to verification, and the best-paid sellers of 2028 will be the ones buyers call to check the machine.

2. Ten agents per seller will make outbound worth less. When every seller has ten agents writing to buyers whose own agents read the mail, unsolicited volume gets filtered at machine speed. The cost of a message falls toward zero. The cost of a real meeting with a real buying group goes up. Anyone pricing on outbound volume is pricing the wrong side of that trade.

3. More content will make you harder to find. Teams producing three to five times more content are not seeing three to five times the results. Buyers’ AI rewards sources that are specific, verified, and consistent. I expect the strongest marketing teams to publish less, prove more, and show up on more shortlists because of it.

4. Procurement, not marketing, will set the new unit of value. Procurement already holds a decision vote in more than half of buying cycles, and it will be the first function to hand comparison work to agents. Activity pricing (per lead, per impression, per seat) is the easiest thing for an agent to line up in a spreadsheet and squeeze. Outcome pricing is the hardest. That asymmetry will reprice whole categories faster than any analyst model expects.

5. The most important marketing hire of 2027 won’t carry a marketing title. The CMO role is already splitting. One half is becoming the owner of what machines know and say about the company. The other half is becoming the owner of the outcome number. The leaders who hold both will keep the job and the budget. The ones holding neither will be reporting to someone who does.

Where your business sits on the Windshield Test

Here is an Uphoff Advisory framework you can download and use as a diagnostic tool in your business. It plots each product line, rather than the company as a whole, on the two variables that decide whether agentic AI is adding value or taking it away.

The vertical axis is what you get paid for, from activity (leads, impressions, content, seats, hours) to outcomes (pipeline, revenue, decisions made). Questions one and three set your height: who is accountable for the result, and whether the customer trusts you for a result or for a tool.

The horizontal axis is whose AI you’re building for, from your own (more output, lower cost) to your buyer’s (being what their AI finds, trusts, and recommends). Question two sets how far right you sit: whether your idea of quality matches what the buyer’s machine rewards.

Download The Windshield Test →

The Rearview Mirror (activity, your AI). AI makes more of what you already sell, faster and cheaper. The buyer’s AI filters the volume and the unit price follows it down. Most content syndication, contact data, and outbound-volume businesses sit here today. Value destroyed.

Margin Without a Moat (outcomes, your AI). You’re paid on results and AI cut your cost to deliver them. The margin is real. The risk is that the shortlist is built before you see the buyer, and efficiency doesn’t get you on it. Value enhanced, for now.

Right Road, Wrong Meter (activity, your buyer’s AI). You show up where buyers’ agents look: verified data, cited research, trusted reviews. You still bill by the lead, the impression, or the seat, so someone else captures the value you create. Value created, not captured.

The Windshield (outcomes, your buyer’s AI). Buyers’ agents rely on your data, proof, and judgment, and you’re paid on the result. Valuation premiums are moving here. Value compounding.

Most businesses will find product lines in more than one quadrant. The same company can be creating and destroying value at the same time, and the P&L will blend the two until the market separates them for you.

What to do with the next AI dollar

If you’re running one of these businesses, split your AI budget into two columns: money that makes your own team more productive, and money that makes you findable, credible, and verifiable to the buyer’s AI. If the first column is most of it, you’re looking in the mirror. Then put a name next to the second column, and a shortlist number next to the name.

If you’re funding one, ask two questions before you look at the AI slide. What unit is on the invoice? Whose AI does the product serve? AI efficiency on an activity unit is margin on a shrinking base. Price it that way.

If you’re building one, move the meter before you scale the product. The step from activity pricing to outcome pricing gets harder with every customer you sign on the old unit.

If you’re starting one, start on the buyer’s side of the machine: proprietary data their agents need and can’t scrape, proof they can verify, and a price tied to the result.

In January I said progress starts by answering three questions honestly. The buyer’s AI isn’t waiting for your answer. It’s giving its own, one shortlist at a time.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

The largest contract Computer Intelligence had ever sold came down to one meeting. I was leading sales and marketing for Ziff-Davis’s research businesses, and on the other side of the table was the president of a large enterprise software company. We gave a long presentation, slide after slide breaking down the pricing. When we finished, he looked at us and said, “Thanks for the lecture on price. Perhaps we can now have a discussion about value.”

We had presented the price as if it were the value. Software companies and services firms still make that mistake every day, and the market has stopped letting them get away with it.

Three in four B2B software companies changed their pricing or packaging in the past year. Most of them changed the model. Almost none of them changed the process that sets the price.

The numbers come from Kyle Poyar’s 2026 State of B2B SaaS and AI Monetization Report, a survey of more than 230 software companies this spring. Hybrid pricing jumped from 25 percent of companies to 37 percent in a single year. Per-seat pricing hangs on mostly at the top of the market: 29 percent of companies above $150 million in ARR still use it. And when the survey asked which pricing model investors would prefer, hybrid was the outright favorite at 35 percent, ahead of outcome-based (26 percent) and usage-based (24 percent). Flat-fee and seat-based barely registered, at 10 percent and 5 percent.

The clearest sign the shift isn’t finished: the single loudest monetization complaint in the survey, especially among flat-fee and seat-based companies, is not enough expansion revenue. That’s exactly the symptom you’d expect when what it costs to deliver the product changed and the price never fully caught up. Even at the earliest stage, when a company has no pricing legacy to unwind, more than a third of companies under $5 million in ARR are still choosing flat fees, a model with no built-in mechanism for expansion revenue at all.

Subscribe now

The gap says more about leadership than about software

None of this is really a software story. I've been brought in as CEO five times, most often to turn around a business in decline or to lead a significant transformation, and pricing was one of the first things I examined. Not because it's a marketing question or a finance question. Pricing is the fastest read there is on whether a leadership team understands its unit economics, and whether it has a real process for setting price to value and managing it over time.

A real pricing process has three parts:

• Market analysis: what competitors, substitutes, and the buyer’s other options cost, including the option of doing nothing.

• Financial analysis: what it costs to win and deliver a unit, and what margin that leaves.

• Positioning: what the outcome is worth to the buyer, and how the offer is framed so they can see it. This is where price turns into value.

Most companies run one of the three and call it a process. Cost-plus shops do the financial analysis and stop. Rate-card followers do the market analysis and stop. Positioning, the part that connects price to value, is the one that gets skipped. It’s exactly what that president was asking for.

Three questions tell you whether the process exists at all:

1. What was the price set against? Cost, a competitor’s rate card, or what the outcome is actually worth to the buyer.

2. What process produced it? All three analyses, run on purpose, with someone accountable for the result, or whichever department had the loudest voice in the room the year it was decided.

3. How is it monitored? A number someone checks every month against current delivery data, or a number everyone quotes from memory that nobody has verified since the day it was picked.

A leadership team that can’t answer all three in one plain sentence each is running a guess it stopped checking a long time ago.

The bigger mistake is usually what you’re pricing, not the price

When I came in as CEO, the pattern I found most often had little to do with the rate. It was the menu. Products were priced à la carte, line by line, when the customer was trying to buy an integrated solution that worked like a recipe. We were selling ingredients to people who came for dinner.

At Thomasnet.com, moving from à la carte to solution pricing lifted our average sale by 40 percent, and the integrated programs performed better for customers than the same products bought one at a time. At Pipeline360, we built a Branded Demand solution that combined programmatic display and data-driven demand generation under one price. Campaign performance improved by more than 35 percent on average, and average order size and customer retention both went up with it.

Neither move was a discount or a price increase. Both decisions weighed unit economics and customer value together: what each sale earned us, and what the customer was actually buying. That's the question the software survey never asks. It tracks how companies charge. It doesn't track whether they're charging for the thing the buyer came for.

À la carte pricing almost always lands in the bottom half of the map below. Every line item gets priced against its own cost, and nobody prices the value of the combination.

Download the Price-to-Value Map and plot your own business on it. Be honest about the vertical axis: it’s the line between a lecture on price and a discussion about value. Most teams that think they price to outcome are pricing to last year’s rate card.

Download The Price-to-Value Map →

Most companies are hiding in one of two quadrants

Flying Blind is the default position for any company that inherited its pricing instead of designing it: no process behind the number, no one checking it, a rate everyone repeats because no one has ever been asked to defend it.

Cost-Plus Discipline looks like the safer place to be. There is a process, and someone reviews the number on a schedule. But if that process is only the financial analysis, running the price against cost instead of what the buyer actually gets, the discipline is rigorously protecting the wrong number.

That quadrant is more crowded than it looks. When the same survey asked what drove pricing on new AI capabilities, the most common answer, at 54 percent, was internal costs and margins. Even the newest thing these companies sell got priced against what it costs to deliver, not what it’s worth.

Value Left on the Table shows up wherever good instinct meets no process: someone in the room already senses the outcome is worth more than the invoice says, but there’s no positioning work or pricing policy to charge for it consistently, so it varies deal to deal and disappears into discounting.

Priced to Value is the only quadrant where the answer to all three diagnostic questions is a current, verified number instead of a guess repeated until it sounded like a fact.

The map covers the first two questions. The third decides whether you stay where you land. A price nobody rechecks drifts back toward Flying Blind, whatever quadrant it started in.

Each quadrant has one move out of it:

• Flying Blind: put a process in place, starting with the financial analysis, so you know what a unit costs to win and deliver before anyone debates what it should earn.

• Cost-Plus Discipline: keep the financial rigor and add the other two parts: the market analysis and the positioning work that turns price into value.

• Value Left on the Table: turn the instinct into a process. Do the positioning work, write the pricing policy, and stop letting each deal renegotiate it.

• Priced to Value: put the number on a monthly review against current delivery data. That’s what keeps it there.

B2B marketing has less excuse than software does for missing this

Agencies and consultancies carry a version of this cost that almost never gets allocated to the deal that required it: discovery calls, a strategy deck built before the contract exists, a campaign concept pitched on spec to win the account. That’s a real cost of winning the business, and most agencies book it as overhead instead of against the account it won. So gross margin can be right in total and wrong for every individual account, and leadership ends up pricing off a number that was never true at the unit level.

And every one of those hours just got cheaper to produce, with agentic AI drafting the first pass of the research, the deck, or the campaign brief. A marketing services business that hasn't repriced against that changed cost is making the same mistake software companies are only now getting measured on, just with a campaign plan instead of a product roadmap.

I built Uphoff Advisory on the bet I laid out in Solo Scale: The New Business Model AI Just Made Possible, that a services business can run priced and structured like a software company. The test above isn’t one I’m suggesting to readers from the sidelines. It’s the one I run against my own business.

The model matters less than the number underneath it

Every business should be checking two numbers every month, side by side. What a unit of whatever it sells earns once every cost of winning and delivering it is counted, not just the ones that show up on an invoice. And what that same unit is worth to the customer who buys it. Per seat, per hour, per project, per outcome, the unit changes. The process that keeps those two numbers in the same conversation doesn’t: market, financials, and positioning, run on a schedule, with someone accountable for the result.

The investors funding your category have already worked out which pricing models reward value. Your buyers are next. Sooner or later, someone across the table will thank you for the lecture on price and ask to talk about value. Have that discussion before they ask, or a competitor will have it with them first.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

That line has been founder gospel for a decade. Scott Adams coined it, and James Clear’s Atomic Habits turned the idea into a movement. It’s also half right.

A goal without a system is hope with a deadline. You know where you’re headed, and nothing on the ground tells you when you’re drifting. A system without a goal fails more quietly: consistent activity that feels like progress and runs straight past the point where it stopped mattering.

The two aren’t rivals. They’re a flywheel. The goal sets the direction. The system checks your position against it. The correction closes the gap. And every correction teaches you something that makes the next one faster. Remove any piece and the wheel stops.

Most of the operating rhythms I’ve inherited as a CEO were missing the correction. OKRs got relabeled “the system.” So did the weekly cadence and the dashboard. None of it changed. It just got a better name. A real system and a goal with a calendar invite look identical in the board deck. They only diverge when the number comes in wrong.

Subscribe now

The test is what happens when the number comes in wrong

One question separates them: when reality diverges from the plan, does someone with the authority to act change course before the next scheduled review, or does everyone wait for the meeting?

If the answer is “we flagged it and we’ll revisit at the QBR,” you don’t have a system. You have a goal with a calendar invite. The rhythm is decorative: the feeling of discipline without the function of it. Does your operating rhythm generate decisions, or reports?

Here is an Uphoff Advisory framework you can download and use as a diagnostic tool in your business.

Download The Correction Test →

I learned this at InformationWeek before I had words for it

I was recruited from Ziff-Davis to CMP Media to take over InformationWeek. It was a $15 million brand, losing money, dead last in a field of seven IT titles on both ad pages and readership.

The goal came first, and we made it impossible to miss: TTL, Take The Lead, on the sign-off of every email, voicemail and meeting. The system came next: four pillars (People, Product, Process, Performance) applied to every function, from sales to editorial to audience development. Not a scoreboard. A diagnostic each team used to find what was holding it back and change what it did next.

Three years later InformationWeek was a $175 million business running at 50% margins and #1 in readership. The goal never got smaller to make the system easier to run. The system existed to serve the goal.

The losers had goals too

Two companies had everything except the correction.

Friendster

• The goal: grow faster than anyone. It had the social network idea years before Facebook.

• The missing system: engineers wanted to fix buckling infrastructure. Leadership kept shipping features, because “grow faster” won every argument.

• The cost: pages took forty seconds to load, and MySpace took the market.

WeWork

• The goal: become the biggest workspace company on earth.

• The missing system: nothing could force a correction when the unit economics broke, and governance kept one person’s desk out of reach.

• The cost: by the 2019 IPO filing, $47 billion in lease obligations sat against $4 billion in committed revenue.

The pattern: both had a goal you could say out loud. Neither built the systems that would force them to notice, and act, when reality stopped agreeing with the plan.

Agents will run the system. Someone still has to design it.

Agentic AI collapses two costs that used to hide a weak operating rhythm: the cost of execution, and the time it takes a signal to reach a decision. When agents do the work continuously, the quarterly check-in becomes an hourly one. The flywheel spins at machine speed whether or not you built a correction into it.

Hand an agent a goal without a system and it will hit the goal. Give a demand-generation agent an MQL target and it will deliver the MQLs. It will also fill your sales team’s calendar with people who were never going to buy, and do it faster than any human team could. That’s Friendster’s mistake, automated.

The systems version gives the same agent a target and a loop: watch MQL-to-opportunity conversion by source, throttle the sources that decay, escalate to a human when conversion drops below a set threshold. The goal stays human. The correction runs around the clock. It’s the loop I described in Loop Engineering applied to the operating model itself.

The market is already sorting this out. Gartner predicts more than 40% of agentic AI projects will be canceled by the end of 2027, citing escalating costs, unclear business value and inadequate risk controls. Read that list again. It describes goals without systems.

When execution gets cheap, the scarce skill is designing the loop: what the agent measures, what it can correct on its own, what threshold pulls in a human, and what pulls the plug. That’s systems-level thinking, and it’s a different muscle than setting ambitious targets. The goal-setters who never built a correction mechanism are about to get exposed at machine speed. The operators who did are about to compound faster than anyone in their market.

Hire and promote the loop designers

You find them in the interview. Ask: walk me through a time a plan you were running turned out to be wrong. Goal-driven candidates describe how hard they pushed to hit the number anyway. Systems thinkers describe what they changed, and how fast.

You keep them by what you reward. Praise someone for hitting a number while ignoring three warning signs, and your best people learn to bury the signal. Reward the one who flagged the problem early and changed course, even when the quarter came in softer.

And you give loop design real weight. Whoever designs correction into your agentic workflows should report to whoever owns the outcome, not sit buried in IT.

Whether you’re building it, running it or funding it

Building it: if your last three “system reviews” ended with “let’s keep an eye on it,” you have goals with better branding.

Running it: give the correction mechanism teeth before you hand the workflow to an agent. Automating a rhythm nobody acts on just multiplies the theater.

Funding it: an operating cadence tells you a founder shows up. Ask what changed the last time the number came in wrong.

Call them goals, systems, OKRs or TTL. None of it works unless something notices when the number comes in wrong and pushes back, whether that’s a person in a QBR or an agent at two in the morning. The goal gives the flywheel direction. The correction gives it speed. Build only one and you don’t have a flywheel. You have a story you tell your board about one.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Gartner’s stock is down roughly 71% from its November 2024 high. Market cap fell from more than $45 billion to about $12 billion. Forrester is worse off: market cap near $105 million, down from roughly $300 million, on a full-year net loss of $119.4 million. These are the two firms B2B buyers have paid for decades to tell them which companies were about to get disrupted. Neither one saw its own name on the list.

Subscribe now

Strategy consulting is cratering at both firms

Gartner’s consulting revenue fell 12.8% in the fourth quarter. Forrester’s strategy consulting bookings dropped more than 50% in a year, and the firm is exiting strategy consulting entirely, alongside an 8% workforce cut announced in February. CEO Gene Hall, on Gartner’s Q4 earnings call, put the cause plainly: “Executives have responded to this by slowing and deferring everything possible. This makes for a much tougher selling environment.”

While research and consulting revenue fell at both firms, Gartner’s Conferences segment grew 11% for the full year, to $645 million. Forrester’s events business, even after cutting event days, is what the firm is redesigning around rather than walking away from. The written word and the paid opinion are getting commoditized inside the very companies built to sell them. The room is not.

Buyers left before it showed up in the balance sheet

Only 13% of B2B buyers used an analyst report in their most recent purchase decision, down 63% since 2022, according to G2’s 2026 Buyer Behavior Report. Eight in ten have sourced a software recommendation directly from ChatGPT or Google’s AI Mode in the last two years. That’s not a preference shift. That’s a category in decline.

The same report found what should worry Gartner and Forrester more than either stock chart: buyers are moving from modeled ROI, a number an analyst estimates from the outside, to observed ROI, proof pulled from their own usage data. The product is starting to make its own business case. The messenger is the part getting cut.

Agentic AI does something more specific here than “get smarter at answering questions.” Gartner’s own July 2026 research puts $234 billion of enterprise software spend at risk from what it calls agentic arbitrage: agents completing a task across systems without a human ever opening the app that used to charge for access. A synthesized report is exactly the kind of task that collapses first under that pressure. It exists to turn public and licensed information into a judgment a human reads. An agent can now do that synthesis in the time it takes to ask the question, at a marginal cost close to zero.

The company that built a model to grade AI disruption never ran it on itself

Forrester spent part of 2026 building its own AI Disruption Model, a tool scoring how exposed other companies’ business models are to AI substitution based on labor intensity, data advantages, and switching costs, among other factors. As far as I can find, Forrester has never pointed that model at itself. It would be the easiest assignment on the team’s desk: one analyst, one afternoon, the same inputs it asks of everyone else.

That silence is its own data point. Gartner’s Magic Quadrant, meanwhile, has carried a credibility tax since long before agentic AI existed. Courts have ruled it isn’t technically pay to play, Gartner won a lawsuit on exactly that claim in 2014, but analyst relations practitioners will tell you privately that meaningful analyst time requires becoming a paying client, which functions the same way even when it isn’t the same thing legally. AI didn’t create that trust problem. It just removed the last reason buyers had to put up with it.

This is the same argument I made three months ago. The numbers just caught up to it

In The Vertical Intelligence Company, I wrote that the cost structure protecting large research firms like Gartner and Forrester had come down, and that agentic AI would be brutal to commodity information while amplifying the value of what it can’t replicate: domain depth, proprietary data, and trusted community. I made that case in June, expecting it to take years to prove out. Even I didn't expect it to take one fiscal quarter.

Watch what they do, not what they say

Both firms are already answering that question with budget, not press releases. Forrester’s own AI product line grew unique users 55% year over year and prompt volume 65%, funded in part by savings from exiting strategy consulting this year.

That’s not a research firm defending its old business. That’s a research firm using the disruptor’s own technology to reinvent itself, on the bet that owning the interface will be worth more than owning the content the interface pulls from.

Neither company is giving up on the room, either. Gartner is expanding it: local C-level communities, longer inquiry days, more destination events on top of a Conferences segment that already grew 11% while research fell. Watching where a company under this much pressure chooses to spend tells you more than anything on its earnings call.

Strategy consulting isn’t an isolated casualty. It’s the leading edge of a category, businesses that sell synthesized judgment as a subscription or a report, that also includes sell-side equity research, credit rating agencies, and most trade press, mine included. None of those industries have been tested at Gartner’s scale yet. All of them are running the same experiment on a slower clock.

I’m not writing this from outside the blast radius

I contribute as an analyst for Cloud Wars. I run an advisory practice that sells judgment for a fee. If the argument here is that commoditized content and standardized opinion are the exposed part of an information business, I’m exposed too, and it would be a thinner piece if I pretended otherwise.

This is exactly what I've argued since the post “Solo Scale”: agentic AI commoditizes general information, and the value shifts to operators who can package decades of earned judgment into an intelligence layer, not a content layer, delivered at software scale and margins. The parts of what I do that read as 'here's my opinion on a trend,' delivered at scale to a broad audience, get commoditized the same way Gartner's research does. The parts built on proprietary access, specific context, and judgment applied to operators actual decisions don't. Gartner's own numbers are drawing that same line around their business right now, whether the company says so or not.

Sort your own information business before the market sorts it for you

Every B2B information, research, or advisory business is really two businesses running under one name: a content layer, and an intelligence layer built on judgment, proprietary access, and work applied to buyer's actual decisions. The content layer scales cheaply, which used to be the entire advantage and is now the entire exposure. The intelligence layer doesn't scale the same way, which used to look like a limitation and now looks like the only part of the business AI can't touch.

To make that split concrete instead of directional, I built the Durability Test. It’s six questions, each scored from one to five, where five means the hardest for AI to replace:

1. What do you actually deliver? A report or ranking is a 1. A live working session is a 5.

2. Where does the underlying material come from? Public information, synthesized, is a 1. Proprietary access nobody else has is a 5.

3. How personalized is the output? The same for every buyer is a 1. Built around one buyer’s situation is a 5.

4. Why does the buyer say they’re paying you? “Tell me what’s happening” is a 1. “Be accountable in the room with me” is a 5.

5. Could a well-prompted AI produce 80% of it in a minute? Yes is a 1. Not without you is a 5.

6. Where does the buyer’s trust sit? In your brand and methodology is a 1. In one person’s track record is a 5.

Download The Durability Test →

Add up your six scores and act, based on the range you land in.

22 to 30: Protect it. 22 to 30: Protect it. You're standing on the intelligence layer of your own business, where AI raises the bar instead of replacing the seat. Use agents to deliver it faster.

13 to 21: Split it. You’re running two businesses under one name. Separate the content lines from the judgment lines this quarter, and move your investment toward the intelligence layer.

6 to 12: Rebuild it. You’re the content layer of your own business, whatever your brand still says on the door. Start with your ten best accounts and ask what each would pay for that a prompt can’t provide.

Gartner and Forrester spent four decades telling other companies where they stood. Now the market is telling them. Run the Durability Test on your own business this week. Gartner and Forrester didn’t get to choose when they found out where they stand. You still do.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

In late 2021, ZoomInfo was worth roughly $25 billion. Today it’s worth about $1 billion. The database didn’t get worse. The customer list didn’t disappear. Something else broke, and every B2B data and intelligence business needs to understand what before the market explains it to them the hard way.

Subscribe now

The company says the moat still holds

ZoomInfo cut its full-year revenue guidance in May, then raised it modestly in August. Net revenue retention has slipped from 90% to 89%: existing customers are generating 11% less revenue than they were a year ago. The stock has fallen from a 2021 peak of $77.35 to under $4.

CEO Henry Schuck’s response has been direct. In May: “ZoomInfo data and our go-to-market context is the ultimate competitive advantage.” In August: “You cannot token your way to building the data asset that we have built.” The first is an argument about where the value sits, data plus context, not data alone. The second is a scarcity claim, and it’s the one most exposed to what’s coming. Whether either holds is the only question that matters here, and it’s not a question the company gets to answer for itself. The market answers it.

I’ve had to answer these questions from the inside, not from a research note

I spent 5 years leading ThomasNet.com, an industrial data and directory business, through the transformation that turned it into a platform Xometry paid roughly $300 million to acquire. Every year of that transformation came down to two questions about our own assets.

The first was scarcity: Could a customer get this data anywhere else, or was it just expensive for someone else to assemble at the time? If it was only expensive, we had a cost advantage, and cost advantages erode.

The second was functionality: Had we built enough into the workflow tools our customers ran their jobs through that they would keep paying even if the data were free? If so, we had a product, not a dataset.

The answers describe very different businesses. The pitch deck reads the same for all of them.

Data moats were built in three eras, and the third removes what the first two stood on

The first era was aggregation: The value was in assembling what was scattered: a directory of industrial suppliers, a database of companies and contacts, a registry of transactions. It took people, time, and capital to build, and that cost was the moat. Customers paid because doing it themselves was not a serious option. ThomasNet began here, as a directory.

The second era was workflow: The data was still scarce or expensive to mine, but now it lived inside software the customer used to do a job: the sales rep’s prospecting screen, the buyer’s supplier search. Switching meant retraining a team and rebuilding a process. Finding another source was the easy part. That is the era ThomasNet’s rebuild into a platform belonged to, and the era ZoomInfo was built for.

The third era is the agentic era: and it attacks both foundations at once. Agents can now assemble facts from public sources for close to nothing, which erodes the cost of aggregation. Customers can use those same agents to build the workflow layer themselves, which erodes the switching cost of the second era.

Schuck says it has started: customers are building interfaces that combine their own data with ZoomInfo’s, though so far only at “the top less than 1% of companies.” That is where a pattern shows up first.

Download The Three Eras of Data Moats →

Agents assemble data and models generate it, and they threaten different businesses

AI reaches a data business by two different routes, and the second gets less attention than it should.

The first is agentic assembly. An agent researching a prospect can crawl the open web, pull public records, and structure what it finds in seconds. Clay lets a customer run a single lookup across many data providers in sequence, which turns a database into one input among several. Nothing gets fabricated. Real facts are collected cheaply. If the capital needed to assemble and refresh a contact database was the real source of the moat, then the moat was a temporary cost advantage all along, and AI removing that cost exposes what the moat was made of. This is the most direct pressure on contact and firmographic data.

The second is synthetic data: records generated by a model with no real-world source at all. It’s gaining steam in training, testing, and modeling, and businesses that sell that kind of data feel it first, because a statistical pattern is exactly what a model can reproduce.

What a model can’t reproduce is a fact about a specific real entity: who owns an asset, what a particular supplier can make, what a transaction cleared at. And it needs real data to stay anchored. A 2024 paper in Nature found that indiscriminate training on model-generated content “causes irreversible defects in the resulting models, in which tails of the original content distribution disappear.” The rare cases go first, and rare cases are where a lot of B2B data earns its price.

Real-world facts can be assembled cheaply but not fabricated, so their value moves to whoever can verify them and stand behind them. I made the upside of that argument in The Vertical Intelligence Company.

Every data business rests on scarcity or functionality, and the agentic era tests both

Scarcity and functionality are the whole test.

Download The Data Moat Test →

Both can be true at once, and the strongest data businesses run on both. Most don’t. Most run on scarcity that was really just cost, dressed up as exclusivity, for as long as cost stayed high. That’s the quadrant ZoomInfo’s own numbers suggest the market has placed it in, Borrowed Time: contact and firmographic data that’s steadily gotten easier for a horizontal AI tool to approximate, sold to customers who are already spending less for it. Schuck’s answer, is a bet that the business can move to the functionality side, that “go-to-market context” is the workflow layer a customer can’t build for themselves. It’s a real bet. It hasn’t been proven yet, and 89% net revenue retention is the market’s current vote on whether it will be.

In the agentic era, the moat is the work you deliver

If the data and the workflow can both be built cheaply, what’s left to build a moat on? The answer is the shift I described in From SaaS to Service as Software. Moats in the agentic era will be built by delivering a service to customers, through your platform or product, that has a material business benefit. Data and software become inputs. The result is what the customer pays for, and someone has to deliver it and stand behind it.

I’ve built a business this way. At Pipeline360 we created the Demand-as-a-Service strategy, a service delivered to customers rather than software they had to run. This model helped drive EBITDA from 25% at inception to 40%.

Moving from selling access to delivering results changes the go-to-market plan in three places, and every vendor with a data asset needs to rethink all three.

• Services strategy. Services stop being the cost center that supports the software and become part of the product. Foundation Capital’s Ashu Garg and Jaya Gupta put it this way after a year of watching AI-native vendors: “In enterprise AI, integration is not a post-sale activity. It is the product surface.” Customers now expect to see the outcome on their own data before they sign. That raises the cost of sale, and vendors have to plan for it rather than fight it.

• Channel strategy. The channel now includes the platforms where the customer’s work happens, and the partners who can carry delivery along with the license. ZoomInfo is building for that reality: its GTM.AI layer is designed to plug into platforms like Salesforce Agentforce and HubSpot Breeze, so the data shows up where the customer’s agents already work. A vendor that insists the customer come to its interface is competing against one that shows up inside the customer’s agent.

• Product architecture. Design the product so customers can build and manage part of the functionality themselves. That feels backwards to any vendor who spent a decade building lock-in. It’s the only way to keep the customers who are going to build anyway. Open the APIs, expose the rules and data as building blocks, and let their agents run on your data and your accountability. The customer who assembles an interface on top of your product is still your customer. The customer who assembles it on top of someone else’s is not.

Run your own business through the test before an investor does it for you

Every founder pitching proprietary data, and every investor who has ever funded that pitch, should be able to answer three questions in one sentence each, without the deck.

1. Scarcity: Can a customer get this data anywhere else, right now? If not, why not? A real answer is structural: a relationship, a regulatory position, or a physical presence.

2. Functionality: If a customer could get the raw data elsewhere, what have you built on top of it that they would miss enough to keep paying for?

3. Result: What measurable business outcome does a customer get from you, and would you stand behind it in the contract?

A strong answer to the first two puts you in The Real Moat. The third answer is what carries that position, or any other, into the agentic era.

The moat was always a question

ZoomInfo hasn’t lost. Not yet. Henry Schuck is making a real argument about where the value in his business actually sits, and it may turn out to be right. But the lesson here isn’t about ZoomInfo. If you run a data business, write down your three answers before your next board meeting, because the market will write them for you if you don’t. If you can’t answer the first two, what you have is inventory. If you can’t answer the third, you’re selling access to a customer who is learning to do without it.

Data built the moats of the first two eras. The work you deliver builds the moat of this one.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

MIT’s NANDA initiative found in 2025 that 95 percent of enterprise generative AI pilots delivered no measurable P&L return, against an estimated 30 to 40 billion dollars invested. Gartner’s forecast for what happens next is worse: more than 40 percent of agentic AI projects will be canceled by the end of 2027, driven in part by escalating costs and unclear business value. The pilots that don’t quietly die become the projects that get killed later.

It would be easy to mistake this data as proof the technology isn’t ready.

It’s not the technology. It’s your business case.

I’ve managed through enough disruptive transformations to recognize this pattern immediately. I watched leaders obsess over MIPS when computing capacity was the scarce resource. I watched them obsess over uptime and query volume during the cloud transition. Now it’s tokens processed and API calls. The instrument changes. The mistake doesn’t: leaders measure the technology. They rarely measure the capability it creates.

Agentic AI didn’t create that dynamic. It just made it faster and more expensive.

Subscribe now

This isn’t a tool you adopt. It’s how you transform your business

I’ve led through two of these technology disruptions as an operator, print to digital and digital to platform, and I’m managing through the third right now: platform to machine, agentic systems that don’t just deliver information, they act on it. The lesson repeats every cycle: the shift looks survivable to the leaders defending the old form, and inevitable to the ones who see the function underneath and build for it early. The leaders who move first aren’t smarter. They just stop asking whether the shift is real and start asking what to build on it.

This one is not a fad, and it is not optional. It will touch every function in business sales, finance, operations, product.

To be clear, because it’s where most companies lose the thread: agentic AI itself was never the point, and the technology wasn’t in the three shifts before it either. The point is what you build with it, and how differently you run the business because it exists.

Confuse the tool with the outcome and you’ll spend two years and a real budget proving nothing at all.

Efficiency is where good intentions go to die

Most agentic AI initiatives get funded on a word that sounds like a standard and isn’t: efficiency. Efficiency is whatever the person reporting it needs it to be that quarter. Hours saved. Tickets closed. Adoption rate. None of it is wrong to track, and none of it survives a P&L review, which is exactly the gap Gartner is describing when it names unclear business value as the top reason these projects get killed.

The only measurement that matters is whether a technology improves your business. Period.

Here’s what that looks like in practice:

Agentic AI success is defined by direct commercial impact: a measurable, attributable improvement in revenue, margin, or cash conversion.

Not adjacent to those numbers. Not a leading indicator of those numbers, someday. Those numbers, named in advance, with a baseline, isolated from everything else moving in the business at the same time.

I learned this discipline decades before a large language model existed. At Pipeline360, EBITDA margin didn’t move because we picked a new technology. It moved because we named the number before we started, built the plan around hitting it, and checked it on a calendar we’d committed to in advance. Agentic AI doesn’t change that discipline. It just raises the cost, and risks, of skipping it.

Think big. Start small. Measure what you can actually prove.

Two instructions, in this order, and most companies only manage one of them.

Think big. The ambition has to match what’s actually happening. Agentic AI will change how your business runs and grows, not which vendor logo sits in your tech stack. If your plan tops out at a chatbot on the website, you’re thinking too small to justify the investment.

Start small. One workflow. One named owner. One proof point, defined before you spend a dollar. The companies pulling ahead right now aren’t running the most pilots. They’re the ones willing to kill nine to get one into production with a number attached to it.

The gap between those two instructions, big ambition and narrow execution, is where most companies fall apart. They either think too small, a tool instead of a transformation, or start too big, a company-wide rollout with no single owner and no defined proof point. What’s missing is a way to know, project by project, exactly where you stand and what has to be true before the next dollar goes in.

The ladder: Assisted, Delegated, Accountable, Compounding

I built a framework for exactly this, a four-stage ladder called ADAC. It doesn’t score your company. It scores the initiative in front of you, and it works the same way whether you’re the one funding it, running it, or diligencing it before you invest.

Download The ADAC Stages

Assisted: a human still owns the task and the result

This is where almost every company starts, and where most never leave. Someone uses an AI copilot for research, drafting, summarizing. Genuinely useful. Not yet a workflow, and not yet anyone’s job to make it one. The sign that you’re stuck here: your only evidence is a usage log. “The team uses it daily” is a habit report, not a business case. Name an owner this quarter, and give them one workflow to rebuild, not ten to experiment with.

Delegated: AI owns the task, a human still owns the result

This is where most of Gartner’s cancellations happen. AI executes a defined piece of work inside a workflow, drafting the follow-up, qualifying the lead, extracting the contract terms, while a human reviews everything before it moves. Real progress, and also where companies get addicted to running pilots instead of finishing one. The tell: a portfolio of pilots, none with an exit date. Freeze new ones. Take the pilot closest to done, name its financial proof point, and get it into production or kill it.

Accountable: AI owns the task and the result, a human owns the exceptions

This is where my sentence stops being aspirational and starts being enforced. The workflow runs end to end. A human steps in only when something breaks, and you should expect an attributable line item on the P&L by now. If you can’t name it, you’re not actually here, you’re still at Delegated wearing this stage’s name tag. The sign you’re stuck: the workflow performs in every test, but leadership still won’t remove the human checkpoint. That’s a governance problem, not a technology one, because nobody has agreed who’s accountable when the agent is wrong on a live customer or on something with a real dollar amount. Name that owner, with defined exposure and a plan to reverse it, before you remove the checkpoint. Not after.

Compounding: AI owns the result and keeps improving how it gets there

The system doesn’t just run the workflow, it improves it, without anyone filing a change request. This is where the mechanism I described in “Loop Engineering” stops being a coding practice and starts being an operating one: judgment encoded well enough that the system holds your standard without you restating it every time. It’s rare, and it’s the only stage where the return compounds instead of holding flat. The sign that you’re stuck: improvement stops the day your one champion gets pulled onto something else, which means you never built the capability, you borrowed one person’s judgment for a while. The fix is institutional: the improvement loop needs a standing role, not a hero.

Score the project, not the company

One warning before you use this stages framework. Don’t score your company. Score the initiative. A business can be Accountable in sales operations and stuck at Assisted in finance, and that’s not a contradiction, that’s normal.

A founder uses this to decide what to fund next. An operator uses it to know where the org actually stands, project by project. An investor uses it in diligence to tell a real deployment from a demo.

1. What’s the metric? Revenue, margin, or cash conversion.

2. What’s the baseline? Where this number stood before the initiative touched it.

3. What’s the required delta? The number this has to move to justify the spend.

4. What’s the date? When you’ll check it.

If you can’t answer all four before you write the check, you’re not funding a workflow. You’re funding a hope.

The Math Isn’t Complicated. The Discipline Is

None of this requires you to become a data scientist, and none of it requires you to wait until the technology settles down before you move. It will not settle down. What it requires is the discipline that has separated companies that make money from companies that make noise since long before agentic AI existed: name the number before you start, know exactly which stage you’re standing in, and refuse to fund the next one until the last one paid for itself.

Confused, frustrated, behind: pick whichever fits this week. None of it disqualifies you from starting. A 95 percent failure rate isn’t evidence you’re late. It’s evidence almost everyone is still measuring the technology instead of the capability it creates. Fix your math first, and you move ahead, not catch up.

Think big. Start small. And never confuse activity with proof.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

”Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Comcast recently filed the closing argument in a 25-year experiment. On June 29, it announced it will spin off NBCUniversal and Sky into a separate public company, unwinding the same combination it built fifteen years ago when it bought NBCU from GE for nearly $40 billion. It’s the second split in six months: Comcast completed the spinoff of its linear cable networks into Versant on January 2. What’s left of Comcast after this one is a pure connectivity company. What’s left of NBCU is a pure content company. Fifteen years and roughly $90 billion in debt later, they’ve concluded the two don’t belong under one roof.

This isn’t a Comcast story. It’s the third chapter in a saga that started with AOL-Time Warner in 2000 and ran through AT&T’s brief, expensive ownership of WarnerMedia.

Content plus pipes has never once worked as a combined business, across three different decades, three different management teams, three different sets of advisors who were certain this time was different.

When a strategy fails identically three times with the smartest people in the industry running it, the problem isn’t execution. It’s the strategy.

I’ve spent thirty-five years on the content side of that equation, running businesses that lived and died by distribution decisions, many of which I didn’t control. If you run a media company, a content brand, or a B2B events business, this isn't background news. It’s the layer you’re standing on. Here’s what I’ve learned watching this pattern repeat: content and pipes aren’t two halves of one business. They’re two businesses optimizing for opposite things, and no org chart can reconcile them.

Subscribe now

Pipes want lock-in. Content wants reach.

A pipe makes money on subscriber lock-in. The value of a cable network or a broadband connection is that customers stay on it, and exclusive content is the tool that keeps them there. Content makes money on reach. The value of a show or a franchise goes up with every additional platform it’s licensed to, every additional audience it reaches, every additional dollar of ad and affiliate revenue that comes with scale.

Put those two assets in the same company and you’ve built a machine fighting itself. The moment content goes over-the-top to compete with Netflix’s global reach, which it has to, it stops protecting the pipe. The moment a pipe tries to hold content exclusive to defend the bundle, it caps that content’s audience at the pipe’s own footprint.

You cannot run the strategy that maximizes the content asset and the strategy that maximizes the distribution asset at the same time. They pull in opposite directions on the same balance sheet.

Wall Street figured this out before the boardrooms did. Pipes get valued like utilities: steady cash flow, bond-like multiples. Content gets valued like growth: optionality, upside, a multiple that assumes the next hit is coming. Bolt them together and the market doesn’t average the two multiples, it discounts both. Comcast’s stock fell 20%-25% over the twelve months before this announcement. That decline is what actually forced the breakup. Strategy followed the stock price, not the other way around.

B2B events are living through the identical fracture

I see this exact tension inside the B2B media industry. Event companies were trading at strong multiples heading into this year. Emerald Holding, a pure-play B2B events operator, had traded in the 12x to 13.7x EBITDA range, well above the 7.81x median for diversified media companies. Apollo took Emerald private at roughly 12x EBITDA in May, closed the deal in July, delisted the stock, and folded it together with Questex into a single combined events platform. The market’s conviction turned into an actual transaction in six weeks. I laid out the data behind that bifurcation in detail in the post: The B2B Event Market Is Splitting in Two. Media-plus-events companies, the combined model a lot of us built our careers on, are trading at a discount to that. Same fracture, different industry.

Events monetize scarcity: a finite room, a badge price, a sponsor paying for proximity to a curated audience. Media monetizes reach: the more free distribution, the better the economics. Bundle the two and the media arm wants to syndicate itself everywhere while the events arm needs to gate access to protect the value of the room. The market prices that tension correctly. It always has.

It was never a pendulum

I’ve told leadership teams for years: you’re either bundling or unbundling, and that’s the whole of business strategy. I still believe that. But I no longer believe it’s a pendulum, swinging back and forth between the two. A pendulum implies bundling and unbundling take turns. They don’t. They’re happening at the same time, on different layers, and mistaking one layer for the whole picture is how experienced operators get caught flat-footed.

At the corporate ownership layer, everything is unbundling: Comcast splitting NBCUniversal and Sky away from its pipes, on a timeline the company still calls on track. Warner Bros. Discovery walked away from its own planned split and agreed to sell itself to Paramount Skydance outright instead, and that sale is now stuck in an antitrust trial that won’t resolve until 2027. Even a deal every advisor involved called strategically obvious can end up fighting over which layer it actually belongs to. At the consumer subscription layer, everything is re-bundling: Disney+, Hulu, and ESPN under one price, Amazon folding entertainment into Prime, streaming fatigue pushing households back toward the mega-bundle they thought they’d escaped. Both are true at once. The question was never bundle or unbundle. It’s which layer you’re standing on when you ask.

Agentic AI is about to move a different layer

This layer doesn't have a multiple yet. The number one complaint from streaming consumers isn’t price, it’s the paradox of choice: too many platforms, too many titles, no reliable way to find what’s actually worth watching tonight. Every streamer has tried to solve this with its own recommendation engine, and every one of those engines has the same conflict of interest. Netflix’s algorithm isn’t built to find you the best show regardless of platform. It’s built to keep you inside Netflix.

An agent with no platform loyalty doesn’t carry that conflict. Its job is resolving your paradox of choice, full stop, even when the answer sends you to a competitor’s service.

That’s a direct transfer of leverage away from the platform and toward whatever sits between the viewer and the decision. We watched this exact playbook run once already: publishers built audiences, then Google and Facebook inserted themselves as the discovery layer and captured the relationship. I wrote about what that actually cost the industry in The Relationship Was the Product: the relationship, not the reach, was the asset that got traded away. This time it isn’t referral traffic and ad inventory on the line. It’s the entire selection and consumption decision.

The center of gravity in user experience is shifting from the streamer to the agent. Whoever solves discovery captures the customer relationship, and none of the incumbent content-plus-pipes players are positioned to own that layer.

That’s Google, Amazon, OpenAI, and Anthropic territory. The next iteration of this 25-year failed formula won’t be a content company buying a pipe company. It’ll be a content company negotiating terms with an agent layer it doesn’t control and can’t acquire. Same trap, new owner of the choke point.

In B2B, that favors events over media: an agent can tell a buyer what to read, but it can't work a room for them. Agents compress the value of content as a discovery mechanism toward zero. They don’t compress the value of presence. They accentuate it.

What this means if you’re running one of these businesses

• Audit which layer you’re actually competing on. If your business depends on being the destination for discovery, you’re exposed. If it depends on being the room, the relationship, or the transaction itself, you’re not, at least not yet.

• Stop diversifying content and distribution into one balance sheet. The math has failed three times at scale. It won’t work the fourth time because your team is smarter than Comcast’s.

• Price for where the multiple is going, not where it is. Today’s events premium and media-events discount are the market using category as a proxy for function, presence versus discovery, and a proxy is never permanent. As agents get better at telling presence from curation inside each category, the multiple will follow the function, not the label. That repricing will move faster than your next planning cycle.

• Build the agent relationship before someone builds it around you. The publishers who lost to Google and Facebook didn’t lose because their content got worse. They lost because they let someone else own the discovery layer. That mistake is available to make again, right now, with agents.

• Separate what’s agent-vulnerable from what’s agent-resistant inside your own portfolio. Discovery and selection are exposed. Presence, trust, and judgment calls made in a room are not. Know which one is paying your bills.

If you don't run a media or events business, ask yourself: what in your business functions as discovery, and is an agent about to sit in front of it?

Download The Agent Exposure Test →

Comcast didn’t fail because the people running it were bad operators. They failed because they were fighting a structural fact with a management chart. The businesses that win the next round won’t be the ones with the biggest combined footprint. They’ll be the ones honest enough to know which layer they’re actually built to defend.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Every technology vendor I have ever sat across the table from was good at exactly one thing: beating the other vendors on the short list. Smarter packaging and pricing. Sharper deck. Faster follow-up. A reference call with the right customer logo. The right relationship at the board level. By the time the deal closed, I knew more about how well they competed than I knew about whether they understood my business. I usually found out the difference during implementation, and I found out for certain at renewal.

That gap is not a sales training problem. It is the whole story of B2B technology for the last fifty years, and almost nobody selling technology today has had a reason to notice it. Until now.

Subscribe now

Customer demand did the selling. Nobody learned to.

Here is the thesis: fifty-plus years of rising corporate technology spending has not materially changed how technology companies position, sell, implement, and renew their products with businesses. The GTM muscle memory built in the mainframe and ERP era is largely still the GTM muscle memory in use today.

That is not because technology companies didn’t have the incentive to get better at it. It is because they never had to. Once the personal computer put a connected terminal on every desk, demand for business technology became so persistent, so self-renewing, that closing a customer stopped being the hard part of the job. Beating the three other vendors chasing the same budget was the hard part. Technology companies got very good at that competition, and very good at telling themselves that winning it was the same thing as understanding the customer.

It never was. It just didn’t matter, because there was always another wave of demand behind the last one to paper over the difference.

We were billed for modules we never used

I’ve seen this move more times than I can count: a competitive pitch dressed up as customer focus. It shows up most often at renewal, the moment the relationship has the least leverage and the vendor has the most.

One renewal stands out. A multi-million dollar contract came up, and the increase landed north of 10 percent. The justification: new fees for modules that were apparently “in the contract” (that we had never used), fees that could be negotiated away, but only if we signed the renewal. From there it was scorched earth. The account team escalated from our CTO straight to my CFO with a thinly veiled threat about what would happen if we looked at a competitor. The rep running the account had never once set foot in our offices.

That is competitive skill working exactly as designed. Nobody on that account understood our business. They understood how to extract a signature under pressure, and for a long time, in a market where demand did most of the selling for them, that worked well.

I’ve had the good fortune of selling in bull markets, and the good fortune of leading businesses through bear markets, where the rules changed and customer focus was the competitive differentiator. Which market I was in never decided that. What showed up the moment the leverage ran out did.

Two skills get confused as one

There are two distinct capabilities buried inside every technology company’s win rate, and the industry has spent fifty years measuring only one of them.

The first is competitive skill: the ability to out-position, out-present, and out-negotiate the other vendors on the shortlist. The second is customer understanding: knowing, with precision, the actual job the buyer’s business needs done, and building a product and a relationship around it. A high win rate can be evidence of either one. Most B2B technology companies have never built a way to tell which.

Plot those two capabilities against each other and you get four quadrants, not two.

Almost every established technology company believes it’s the Category Leader because its win rate has always been strong. Most of them have actually been the Borrowed Win, protected the entire time by demand strong enough to make the difference invisible.

Download The Borrowed Win Quadrant →

What was hiding the difference

The evidence that the Borrowed Win is more crowded than anyone wants to admit is already sitting in the data. Gartner has found that 60 percent of technology buyers involved in a renewal decision regret nearly every purchase they’ve made. McKinsey’s own research found that only 16 percent of digital transformations both improved performance and sustained that improvement long term. Those are not implementation statistics. They are customer-understanding statistics, showing up two steps downstream from where the deal was actually won.

A vendor that understood the job would not produce those numbers at that scale, for that long, across an entire industry. A vendor that was simply good at winning competitive bake-offs would produce exactly this. The industry has had the data in front of it for years and read it as a services problem instead of what it actually is.

Why the cover is coming off now

The demand cushion depended on one structural fact: the buyer’s evaluation ran through a human, and a human can be out-positioned, flattered, and outlasted. Agentic AI removes that fact, not gradually, but as fast as agents get adopted inside procurement functions.

I’ve written before about what happens when the buyer’s shortlist gets built by an agent before a rep is ever involved, in “The Buyer Left the Building”. An agent evaluating vendors on behalf of a buyer does not care how the deal was packaged or priced. It does not respond to a sharper deck. It does not remember a good reference call. It checks whether the product demonstrably does the job it was asked to evaluate, and it moves on if the answer is unclear. Competitive skill, the entire capability the industry spent fifty years perfecting, does not register on that test. Customer understanding is the only thing that does.

That shift lands hardest on the people who can least afford it. I’ve also written about the asymmetry between the rep who walks away from a bad decision and the buyer who staked their internal credibility on it in ”The Decision Was Made Without You”. That buyer is exactly who is now handing evaluation work to an agent built to catch the gap between a vendor’s pitch and a vendor’s actual fit, because they were the one left holding the outcome the last time the pitch won and the fit didn’t. The Borrowed Win quadrant only worked because the buyer couldn't see the difference before signing. The agent can.

What this actually requires

• If you sell technology: Stop measuring win-loss against the competition and start measuring it against whether the customer’s job actually got done, twelve months out, not thirty days after signature. Treat a strong win rate with no customer-understanding evidence behind it as a warning, not a victory lap. Rebuild the account team around the person inside the buying business who is personally exposed by the decision, not the procurement function that approved the budget, because that is who is about to be doing the evaluating with an agent at their side.

• If you run a business: Stop running evaluations that reward the vendor who is best at being evaluated. Do the work of defining the job you actually need done, with the same rigor you’d apply to a hire, before a single vendor is in the room. Hold every vendor accountable to the outcome past the signature, not just the promise before it, and treat renewal, not the initial close, as the real test of whether the fit was ever real.

Fifty years of demand did what strong demand always does. It let a lot of companies mistake being hard to say no to for being right for the job. Agentic AI does not care how hard you are to say no to. It only checks the job.

You have spent fifty years selling technology. Whether you ever understood your customer's business is what agentic AI is about to find out

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory: a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

The board deck is due. The forecast came in soft a week ago, and everyone in the building already knows it. Nobody has said so out loud, because the first person to say it risks sounding like they never believed in the plan.

That’s the position every CEO eventually finds themselves in. Above you, a board and investors who signed off on a number and are anxious to know how the business is performing. Below you, a management team that either doesn’t know how bad it is yet or is too nervous to bring it to you first. Both conversations run through the same person: you. And most days, you're the only one standing there.

A recent Harvard Business Review piece on delivering bad earnings news got the mechanics right: say it early, quantify the shortfall in concrete terms, kill the speculation before it starts, pair the bad news with a plan. All good advice. It also assumes you already know two things that are actually the hard part: what you told people to expect in the first place, and what really happened, as opposed to the version you’re telling yourself happened.

In five CEO roles, I’ve missed numbers and exceeded them. Thankfully, while I was more often on the positive side, the misses taught me far more. What follows are the systems those misses forced me to build: how I manage a business against its budget, how to diagnose the causes of a shortfall, how to build a bridge plan and how to communicate about performance with the board and your operating team. The charts and frameworks that come out of it are yours to download and use in your own business.

Subscribe now

Most misses get decided before the year starts

Ask ten CEOs whether their team can tell the difference between the budget, the forecast, the pipeline, and the quota assignment. In my experience, most can't. Those are four different numbers doing four different jobs, and most organizations flatten them into a single phrase: the number. That flattening isn’t harmless. It creates confusion at the leadership table, produces inaccurate numbers, breeds anxiety that has nothing to do with actual performance, and every so often, sets up a mistake serious enough to be catastrophic.

The budget is what the board approved, the number the business is actually run on, the baseline every operating decision gets measured against all year. The forecast is what you actually expect to deliver in the month, quarter and year, updated as real data comes in, and it’s usually the number the board cares most about, even though budget is the one they signed off on. Pipeline is the raw material underneath the forecast, always larger than what will actually close, discounted by stage and probability. The quota assignment is what you’ve asked the sales organization to carry, often set above the forecast on purpose as a stretch, and it should never be confused with what the company expects to report.

The Four Numbers: budget, forecast, pipeline, and quota assignment, mapped by owner, what each measures, how often it’s updated, and who sees it.

Download The Four Numbers →

Confuse those four numbers and a casual comment turns into a real problem. An executive tells a board member in the hallway that the team is running 10 percent ahead of the number, and everyone nods. That’s a good update if the number in question is budget. If it’s forecast, and the forecast has been trending below budget for six weeks, that same comment is wrong, and by the time it gets corrected, the board has already recalibrated around a number that was never real. The discipline isn’t more communication. It’s specifying which of the four numbers you’re talking about, every single time.

The tool I've used for thirty years to manage the forecast is a Best Case, Worst Case, Realistic Case model, built specifically for quarterly and annual forecasting: not a single figure, but a range with a 10 to 15 percent variance band that tracks actual and projected performance against budget as the quarter unfolds, not after it closes. Run the business against that range every week and a shortfall shows up as a trend line long before it shows up as a surprise, which is what actually lets you act on it early. The board benefit is real, but it's a byproduct: a board that has already seen the worst case in writing doesn't treat a shortfall as a betrayal.

Download The Three-Case Model →

InformationWeek: $15 million and losing money to $175 million and $75 million in EBITDA, in four years. We built the plan on a series of key indicators, not the year-end target: qualified pipeline by vertical, ad page growth against the prior year, market share against 6 competitors and market share by individual advertisers. Every leader could recite them, so nobody was surprised by the outcome. Growth that fast creates its own pressure. Systems built for a $15 million business buckle at $75 million, and the leader who gets a business from one tier to the next isn’t always the leader who gets it to the tier after that. Scaling meant rebuilding systems, changing out people who weren’t scaling with us, and resetting expectations upward every quarter, because a board that just watched you beat the number by 20 percent won’t accept a repeat of last year’s target.

The Hollywood Reporter: 30 percent growth in a single year, against a legacy print category everyone assumed was done growing. The number came from a specific plan built around Total Reach, an early, integrated print and digital offering, not a guess. We passed Variety in share of market for the first time, and the board never had to ask what was driving it, because they’d seen the components every quarter.

Dramatic growth curves are the trickiest forecast of all: no history to build a plan from. Game Developer Conference and Black Hat both grew more than 45 percent a year, on plans built from real registration and sponsorship data, not hope.

The pattern I’ve watched repeat: the target got set to satisfy a board conversation, not because anyone had evidence it was achievable. A recent survey of more than 2,000 business leaders found 81 percent of companies missed their sales forecast in at least one quarter over a two-year stretch. Most misses aren’t a market surprise. They’re the plan finding out, in public, that it was never that precise to begin with.

Don’t explain it before you understand it

The moment the number turns negative, there are two instincts, and both are wrong. The first is to defend the plan: the market’s soft, the timing’s off, give it another quarter. The second is to overcorrect: blow up the strategy, replace the team, apologize your way through the next board meeting. Both are a way of deciding what happened before you’ve actually found out.

I give myself, and I'd tell any operator to give themselves, roughly a week before forming an opinion out loud. Not a week of silence. A week of gathering the actual signal: pipeline conversion by stage, not just pipeline size. Win-loss detail on the specific deals that slipped, not a summary someone else wrote for you. What customers say when I get them on the phone myself, not the paraphrase in a deal review. What's actually changed with competitors, a price move, a new feature, a new entrant, and whether they're seeing the same softness we are. Whether customers are telling us directly that budgets are under pressure, which is a market signal, not an excuse. Whether the shortfall is broad across the business or concentrated in one product, one region, one team. The diagnosis takes discipline precisely because the pressure to have an answer arrives faster than the evidence does.

This is where most leadership teams skip a step. They jump straight from “we missed” to “here’s the plan to fix it,” without ever confirming what actually broke. A recovery plan built on the wrong diagnosis doesn’t just fail to fix the problem. It burns the credibility you’ll need the next time something goes wrong, and there will be a next time.

Every miss traces back to one of four places

After thirty-five years of this, almost every miss I’ve seen sorts into one of four categories, and the category determines nearly everything about what you do next.

Download The Miss Diagnostic →

Miscalculation is ours, and it was wrong from the start. The gap shows up in month one, not month nine, because the leading indicators were off before you ever went to market. This is the most common cause behind a new launch that misses: not a market that turned, but a target nobody had real evidence for. The fix is to rebuild the model instead of defending it. Reforecast off real data, set the new number, and hold it.

Market dynamic the target was fair when you set it but the market has contracted. The whole category slows at once, and peers are missing similar numbers. This is what happened across most B2B print brands after the internet arrived faster and cut deeper than any of us modeled: the plan wasn’t naive, the ground underneath it moved. The move here is to adjust the timeline, not the thesis. Protect the plan’s core assumptions and reset the pace against them.

Execution gap is ours, and the target was right. Pipeline, hiring, or launch dates slipped against a plan nobody disputes. The math still works. This is the one that’s hardest to admit, because it’s the one closest to home, and the fix is operating discipline: name what didn’t get done, who owns it, and the date it gets done by.

Competitive gap Our competitive position moved under us. You’re losing competitive deals you used to win. Win rate is down against a market that isn’t shrinking, which is a different challenge than a market that is. The fix is to fund the product gap directly. No amount of sales effort outruns an uncompetitive offer.

The distinction matters because each of these four demands a completely different conversation with the board and a completely different set of actions with your executive team. Mistake an execution gap for a market dynamic, and you’ll spend two quarters waiting for a recovery the market was never going to hand you. Mistake a market dynamic for an execution gap, and you’ll fire people for a problem that was never theirs to own.

A shortfall should never arrive without a plan attached

At every company I’ve run, the rule was simple: if you’re forecasting a shortfall, you don’t bring me the problem alone. You bring a Bridge Plan, a specific plan to close the gap and get back to or above budget within the next two quarters, with the specifics, the owner, and the dates attached. Not a general promise to work harder. An actual bridge from where the number is heading to where it needs to land.

Download The Bridge Plan →

This does two things a recovery conversation after the fact never can. It surfaces the shortfall while it’s still fixable, months before it would otherwise show up in a board deck. And it forces the person closest to the problem to do their own version of the diagnosis above (miscalculation, market, execution, competitive) before it ever reaches me, because nobody can build a credible bridge without knowing which one they’re actually standing on.

The board and the team need two different conversations, not one memo

The board wants the diagnosis, the revised trajectory, and the plan, in that order, delivered before they have to ask. State the shortfall in dollars and percent against the number they approved, not a description of how things feel. Tell them which of the four categories you’ve landed on and why, with the evidence, not the instinct. Give them a revised number with a date attached, and be available for the follow-up questions instead of hiding behind a deck. What kills credibility with a board isn’t the miss. It’s finding out later that you knew more than you said when you said it.

The team needs something different: clarity about what’s actually being asked of them now, and honest signal about whether this is about them or not. If the diagnosis says market dynamic or miscalculation, say so plainly and protect the team from carrying blame that isn’t theirs. If it’s an execution gap, say that plainly too, with names, owners, and dates, because a team that already knows something went wrong respects a direct conversation far more than a vague one. Silence in front of a team gets read as either panic or indifference. Neither is true, and both are worse than whatever you’re actually feeling.

What actually happens in the next thirty days

Get the real number before you get an opinion: A week of diagnosis before any external conversation, board included.

Diagnose before you communicate: Know which of the four categories you’re in before you draft a single sentence to anyone outside the leadership team.

Ask whoever owns the shortfall for a Bridge Plan, not an explanation: A specific plan back to or above the number within two quarters, with an owner and a date, not a promise to try harder.

Tell the board first, in writing, before the meeting: State the shortfall in dollars and percent against the number they approved, not how things feel.

Match the message to the diagnosis, not to your mood: A market dynamic gets a timeline reset. An execution gap gets names, owners, and dates. Don’t apologize for a market you don’t control, and don’t excuse a team that didn’t do its job.

Tell the team the same week, not after the board reacts: A delay reads as evidence you don’t trust them with the truth, which is a worse problem than the miss itself.

Put a recheck on the calendar before you leave the room: Thirty, sixty, ninety days out, tied to the same leading indicators that should have flagged this the first time. If you don’t have those indicators yet, building them is the actual fix, not the recovery plan itself.

The number was never the hardest part

Go back to the room I opened with. The board deck is due, the number is soft, and nobody has said it out loud yet. The instinct in that room is to treat silence as safety. It isn’t. Every day that passes without confronting the truth and diagnosing the problem is a day the board and the team are each writing their own version of what happened, and neither version is likely to be generous.

I’ve missed numbers plenty of times across five CEO roles: print brands that declined faster than the plan accounted for, launches where the market taught us something the projections couldn’t have. None of those misses cost me what a bad diagnosis or a slow, evasive conversation would have. The number is never actually the hardest part of the job. Deciding what it means, and saying so before someone else decides for you, is.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

“Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive.

I was a newly minted VP at Ziff-Davis, in a job that felt way over my head, and likely was. With my one-year-old daughter asleep on my chest, I started reading “What It Takes”. I thought I was reading a book about a presidential primary. I didn’t realize I was reading a book that would help me identify a massive challenge I was facing: it told me I was built for business leadership. It also warned me something would break if I didn’t get smarter about managing it.

Richard Ben Cramer’s “What It Takes: The Way to the White House” is about six men who wanted to be president. My daughter Olivia was one year old, and somewhere in the middle of that brick of a book I stopped reading it as political history and started reading it as a mirror.

Cramer didn’t write about policy. He wrote about appetite: what makes six people volunteer for the worst job interview in the world, eighteen months of having their entire life picked apart in public, on the chance they get to do something enormous. Thirty-four years later that’s still the only question in the book worth remembering.

Subscribe now

“Crazy enough” was the easy take away. The harder one took longer to learn.

Some part of me already suspected I had the temperament for what I was stepping into. I didn’t shrink from decisions made without enough information, and I wasn’t especially afraid of being wrong in front of people who could make my life difficult about it. Cramer’s book confirmed that part quickly. Six people, three Republicans effectively down to two, four Democrats, all of them certifiably a little crazy in the same specific way, willing to trade a normal life for a shot at an impossible one.

The second lesson took longer, and it was the one that actually mattered. Wanting a job badly enough to survive it and knowing how to carry it without dropping everything else are not the same skill. I had the first one. I was still working on the second, and the evidence was standing right in front of me every night: a business demanding more of me than I knew how to give it efficiently, a marriage that needed attention I wasn’t always paying, a daughter who wouldn’t be a one year old twice. Something was going to break if I kept running leadership on instinct instead of a system.

I still get some version of the same question every few months, usually from someone weighing a move into management who wants me to tell them whether they should take it. My answer hasn't changed since I first heard myself say it out loud: be careful what you wish for.

The question deserves a real answer, not a warning

That line was never meant to talk anyone out of leadership. It was meant to make them actually answer the question instead of just wanting the title. Most people who ask are conflating two different things: wanting the exposure and status that comes with the title; and being built for the actual mechanics of doing the job, deciding without complete information and getting judged in public for the decision either way. Those are separable, and running one against the other tells you more than either one alone.

Plot tolerance for ambiguity, how you decide before the picture is complete, against appetite for exposure, how you hold up being watched, wrong, and reassessed in public, and four seats show up, not one.

Download Do You Actually Want the Seat? →

High on both is the Candidate quadrant. Cramer's six all lived there, not because they were the smartest people in the race, but because they could stand the room. Low ambiguity tolerance paired with high appetite for exposure is the Performer: commanding once the plan is set, exposed the moment it isn't. High ambiguity tolerance paired with low appetite for exposure is the Strategist: seeing three moves ahead and genuinely fine not knowing how it resolves, as long as their name isn't the one attached to the outcome in public. Low on both is the Specialist: doing the org's best work against a clear brief, someone for whom both ambiguity and scrutiny drain rather than sharpen.

I landed in the Candidate quadrant early in my career, and that told me almost nothing useful by itself. It told me I could survive the job. It didn't tell me I knew how to run it without a system underneath me, and that gap nearly caught up with me.

The scrutiny used to have an off switch. Now it doesn't

Every one of Cramer’s six candidates got a concession speech. Win or lose, a day arrived when the campaign ended, the coverage moved on, and the verdict closed. Business leadership used to run on its own quieter version of that system. A bad quarter got explained away by the next earnings call. A rough town hall got buried under Monday’s news cycle. A 360 review happened once a year, on a schedule a leader could see coming and prepare for. The evaluation had a rhythm, and the rhythm had recovery time built into it.

That rhythm is gone, and it didn’t fade out gradually. I’ve argued that this same cycle used to exist on the sales side of B2B, a human somewhere in the loop who could feel when something was off and course-correct before it became “a lost deal nobody could explain”.

Agentic AI is doing the identical thing to how leaders themselves get evaluated. An agent summarizing a week of all-hands transcripts doesn't wait for the quarterly engagement survey to catch the gap between what a leader said and what the room actually heard. A sentiment tool reading internal Slack doesn’t need a listening tour scheduled before it clocks that morale moved. A board member watching a live dashboard doesn’t need the deck to ask why a number shifted in the last 48 hours. Ground truth doesn’t wait for the annual review anymore. It updates continuously, and it doesn’t ask permission first.

The key take away here isn’t more monitoring. It’s the disappearance of the recovery window. Episodic scrutiny gave a leader a window to have a bad week and fix it before the bad week became the story. Continuous scrutiny shrinks that window to almost nothing. Every live signal is a small, ongoing referendum, and the referendum never adjourns long enough for anyone to catch their breath.

That changes what "crazy enough" has to mean now. His candidates needed the temperament to survive one campaign with a clear end date and a known finish line. The same temperament today has to run several overlapping evaluations at once, on a clock that never calls a recess.

The same six traits, every one of them wearing a different shape

Everyone who's asked me some version of "should I take this job" over the last three decades was really asking about six things: resilience, tolerance for ambiguity, the discipline to let some fires burn so you can fight the ones that matter, the ability to take and carry real risk, optimism, and whether they had anyone around them who could catch what they couldn't see themselves. None of the six are new. Here's what continuous scrutiny did to each one:

1. Resilience used to be measured in quarters, a bad stretch absorbed by a good one three months later. Continuous sentiment and performance signal collapses that runway to almost nothing. A leader’s response to one difficult week is now legible in near real time instead of filed away until the next board deck, and the recovery has to happen on the same compressed clock the scrutiny runs on.

2. Tolerance for ambiguity used to mean the data wasn't in yet, and the skill was patience, knowing how long to wait before you moved. More often now the data is in, constantly, streaming from a dozen live sources at once, and the harder skill is deciding while the picture keeps moving underneath you instead of holding still long enough to read it.

3. Letting fires burn gets harder before it gets easier. Agents surface more signal, not less, more anomalies flagged, more decisions that technically require a human sign-off. The skill shifts from knowing which fires to walk past to building a real triage system for the fires a machine hands you before you’ve noticed the smoke yourself.

4. Managing risk now includes a category Cramer’s six never had to deal with: the exposure created by handing judgment to a system nobody in the room can fully audit in real time, and answering for its calls as if they were your own, because to the board and the market, they are.

5. Optimism has to stay grounded instead of promotional, because the promotional version doesn’t survive contact with a dashboard that updates hourly. The clearest recent example wasn't a founder's keynote, it was three departing CEOs telling their boards the honest version of the same sentence. As I wrote when Doug McMillon of Walmart, James Quincey of Coca-Cola, and Shantanu Narayen of Adobe all made a version of the same call within months of each other, the real optimism in that moment wasn't "I can do this." It was "this is worth doing, whether or not I'm the one who finishes it."

6. Support systems is the trait I have the least patience for people skipping, mostly because it's the one I skipped longest myself, and it's the trait continuous scrutiny punishes fastest. It covers more than having someone to call when things get hard. It means a leadership team strong enough to catch what you can't see about yourself, a management structure that keeps running when you're not the one holding it up, and enough deliberate structure at home to keep a marriage and a family from surviving on whatever attention is left over at the end of the day. A leader with nowhere to process a hard week, at the office or at the kitchen table, before the next signal lands starts making decisions from fatigue instead of judgment.

The trait I nearly failed

Here's what actually would have broken early in my career if I hadn't taken Cramer's second verdict seriously. Not the business. I was resilient enough, ambitious enough, comfortable enough with ambiguity to keep the job moving. What was at risk was the husband and the father fully showing up in addition to the business leader, and I didn’t fix that by trying harder. I fixed it by building actual structure around how I ran my time and my attention, instead of assuming strength of will alone would cover the gap.

The data on this hasn’t gotten friendlier. Somewhere between 55 and 60 percent of CEOs report feeling isolated at the job they fought to get. Nearly all of them say they’d welcome coaching or outside counsel. Roughly two-thirds get none. New leaders have it worse. More than seven in ten say they've felt lonelier in the early stretch of a new role than they did before they stepped into it. The instinct that gets rewarded on the way up, carry it yourself, is precisely the instinct that has to be unlearned once you’re actually in the seat.

I didn't build a peer group or hire a coach early in my career. I didn’t have the language for what I needed yet. What I did have was a book that scared me into taking the second lesson as seriously as the first, and enough sense to start building something more deliberate than instinct before our home life was significantly impacted.

Three of the four quadrants are leadership too

Not everyone belongs in the Candidate quadrant, and the other three quadrants were never the lesser answer. The Performer runs the room once the plan is set and deserves a real system around them instead of a lecture about toughening up. The Strategist is often the best chief of staff or strategy chief a Candidate will ever have, not a leader who fell short. The Specialist does work most organizations can’t survive without, and the next move from there should be toward mastery, not toward a title that was never the right fit to begin with.

“Be careful what you wish for” was never a warning to stay small. It was an instruction to find out which of these four quadrants you’re actually built for before you take the promotion, not after you’re already in the job and something at home is starting to give.

Olivia is all grown up now with children of her own, and the book is still on my shelf. Prioritizing my family didn't cost the business anything. It made me a better leader, and it made those years worth living, not just surviving. What I remember about the book isn't the six candidates. It's the young business leader holding his daughter, finding out the hard way that wanting the job and knowing how to carry it are two different tests, and that you had better pass both.

Keep going: What Are Employers Actually Hiring For Right Now? and Never Make Them Feel Stupid

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

”Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

Open your LinkedIn feed and read ten thought leadership posts in a row. Count how many you could tell apart if the names were stripped off. That number, not how many posts got published this year, is the real measure of how broken B2B content marketing has become.

Subscribe now

Enterprise technology went mainstream. So did the people who ran it.

In the late 1980s and early 1990s, enterprise technology was a real and growing market, almost entirely invisible outside the people who bought it. That changed once the personal computer, then the internet, went mainstream, and companies like Microsoft, Intel, Cisco, Oracle, and SAP rode the same wave into the business pages, their stocks bid up on the size of the opportunity ahead.

Mainstream companies produce mainstream curiosity about the people running them. It became routine for a reporter to ask Bill Gates, or Andy Grove, or Steve Jobs what they thought about the next wave of technology, and tech leaders turned into public figures in their own right, writing books and giving speeches that had nothing to do with their product line. As the internet handed every company its own printing press, these same companies started publishing directly to an audience for the first time. That instinct, not a marketing department’s idea, is where Brand as Publisher actually started.

B2B media was dying at the exact moment vendors learned to publish.

The other half of this story was happening inside the B2B Media industry. The internet that was making tech companies mainstream was disrupting the high-margin print business B2B media had run for decades. One response was custom publishing: vendor-sponsored content, clearly labeled as such, running inside the same magazines and web sites that carried the independent coverage. At some point, Brand as Publisher and custom publishing converged, and the modern thought leadership and content marketing era was born from that collision.

Much of what came out of that convergence worked well. Buyers trusted content that came directly from the companies building the technology, and used it to evaluate and shortlist vendors. This shouldn’t have been a surprise: Oracle, Salesforce, SAP, IBM, and others hired the B2B tech journalists that declining media companies could no longer afford, and put them to work producing exactly the kind of content they used to produce for media companies. So far, so good. Then marketing automation arrived, and the math changed.

I was one of the pioneers of this, and I had plenty of company. I did not see what was coming next

I’ve spent over half my career running B2B media businesses, and I was one of the pioneers of the very things I am describing here, and so were plenty of my peers in B2B media. Sponsored media, custom publishing, marketing automation, demand generation, and thought leadership at CMP Media and UBM TechWeb, the same adaptation a lot of B2B media publishers were making as the internet ate our print businesses out from under us.

I was one of the pioneers of this shift, at CMP Media and later UBM TechWeb. In an effort to evolve a business model the Internet was steadily disrupting, B2B media companies used their traditional editorial processes to build sponsored content operations, and applied the new technology of marketing automation to open a new revenue line in demand generation. In 2010, while CEO of UBM TechWeb, I wrote a white paper called “Marketing-as-a-Service” that outlined the path forward. This was before “content marketing” or “brand as publisher” had become common terms. That model is what eventually got operationalized by brands, custom content providers and demand generation companies, dramatically escalating the volume of content being produced.

What we didn’t anticipate was how fragile that model actually was, or where the talent would go once it started cracking. Bob Evans, my editorial partner at InformationWeek and UBM TechWeb, became Chief Communications Officer at SAP, then Oracle, and now runs Cloud Wars. Rob Preston, Fritz Nelson, Stephanie Stahl, and Tom Smith, editors I worked alongside at InformationWeek and UBM TechWeb, now lead marketing and content organizations at some of the largest technology companies in the world. One of them now runs the Content Marketing Institute. A flood of talented B2B editors moved from declining media companies into the vendors they used to cover, carrying real editorial skills and judgement into organizations that would eventually systematize it into something else.

I made this argument in May. AI just put it on fast-forward

In May, I argued that B2B marketing had substituted measurement for the work itself, that attribution models and MQL dashboards had become a stand-in for brand, positioning, and judgment “How B2B Marketing Lost Its Way”. I ended that post with a prediction that is proving out faster than I expected: what survives in an agentic world is genuine brand authority, distinctive positioning, authentic thought leadership. The things that cannot be tracked are the things that matter most.

As Generative AI, marketing automation, and every tool in between scaled thought leadership and content marketing output exponentially, what got optimized was scale, not judgment, taste, or any real understanding of where a piece of content actually supported a purchase process. Fifteen years of tools were built to measure clicks, downloads, volume, and SEO score. None of them were built to measure whether a reader trusted what they read, or whether that trust moved them one step closer to a decision.

The tools did exactly what they were built to do. That was the problem.

Marketing automation was the opening move. What it triggered next was an arms race. I already tallied where that race ended up in "The Stack That Broke Marketing": 15,384 commercial martech tools as of last year, up nearly a hundredfold since 2011. Utilization sits at 49 percent, meaning close to half of everything B2B marketing spends on technology produces no active output. Content and thought leadership inherited the same math: more tools bought than anyone had a plan to use.

Fifty-two percent of newly published articles on the web were AI-generated as of last year, up from near zero four years earlier, per Graphite’s analysis of 65,000 English-language articles. That number alone would be a footnote if the content underneath it were still doing its job. It is not. Ninety-five percent of B2B marketers now use AI tools in content production. Fewer than four in ten report a measurable performance improvement from them. Ninety-six percent of B2B brands produce thought leadership. Only somewhere between four and eleven percent rate their own program as advanced.

Fifty-six percent of B2B marketers cannot attribute ROI to their content at all. Marketing budgets grew nearly nine percent heading into this year anyway. Traffic volume gave content marketers cover, so a real ROI conversation never had to happen.

Reading these numbers a troubling pattern emerges. Adoption is universal. Effectiveness is not. That gap is not a rollout problem better prompting fixes next quarter. It is what happens when a discipline built on editorial judgment gets handed to a machine built to optimize what a dashboard can report, run by people who never had to answer for the one thing a dashboard cannot show: whether anyone trusted what they read.

Buyers noticed before the metrics did

Buyers do not have that blind spot. A Gartner survey published this year found that sixty-nine percent of B2B buyers still turn to a human sales rep to validate AI-generated insights before they trust them, even as those same buyers say they prefer a self-service, rep-free experience. Content marketing was invented, in part, to let a company scale trust without a person in every conversation. The tools-first version of it did not scale trust. It just made buyers trust the human relationship more.

An Outsell study of 226 B2B buyers found reliance on vendor-produced content falling across the board, down fourteen percent among the end users closest to a purchase decision, while reliance on peer communities, analyst reports, and trade press rose by double digits over the same period. I used this same data two weeks ago to make a case about how AI changed the way B2B buyers evaluate vendors. Read from the content side instead of the buying side, it makes an even sharper case against the content itself.

Buyers are not abandoning research. They are relocating their trust to sources a company cannot buy its way into. None of this is new: Edelman’s original thought leadership research found fifty-six percent of B2B decision makers got no value from what they read, and eighty-six percent rated it merely good, mediocre, or poor. That predates generative AI by close to a decade. The tools did not create this problem. They gave it a much faster printing press.

I asked Tonia Ries, who spent a decade working on Edelman's Trust Barometer research, what she's seeing now. Here is her take:

“We’ve spent years measuring what audiences do with content: whether they click it, share it, or respond to it, treating those behaviors as proxies for trust. But attention is not the same as trust”.

“Audiences are reading the trust signals in the content itself. Does the company demonstrate real competence? Is it candid about complexity and uncertainty? Does it reflect the reader’s interests as well as the company’s? Does it give a potential customer a reason to rely on the organization when their career is at stake?”

“AI is making content dramatically easier to produce, which means those trust signals are going to become even more essential. If your content is capturing attention without also earning trust, your funnel will start leaking, no matter how healthy it looks from the top”.

Tonia Ries, former Edelman executive

Run the audit before your buyers run it for you

Here is a way to determine where your own programs stand, before a reader does it for you. I’m calling it the Narrative Index.

Take your last three published pieces and score each one, one to five, against these seven questions:

1. Byline accountability. Does a named person's judgment stand behind this, someone accountable if it's wrong?

2. Source diversity. Does it draw on more than one voice or dataset, or is it a single synthesized take dressed up as insight?

3. Distinctive point of view. Could a competitor publish this exact piece with their logo swapped in?

4. Third-party earned trust. Has anyone outside the company chosen to cite, share, or reference it, unprompted?

5. Tool dependency. Did this require a human editorial decision beyond hitting publish?

6. Incentive alignment. Is the team that made this rewarded for whether it changed a reader’s mind, or for how many of these shipped this month?

7. Job-to-be-done fit. Does this move the buyer forward on the job they hired your category to do, or does it just talk about the category?

Add the seven scores. Thirty-five is the ceiling. Twenty-nine to thirty-five means your program is still narrative-led, rare enough to protect rather than optimize further. Eighteen to twenty-eight means you are at risk, which is most programs, and still salvageable with real editorial judgment back in charge. Below eighteen, the tools are running your program, and your content will continue to decline in impact.

Download The Narrative Index →

The test isn’t rhetorical. Run it on your own byline

Go back to your feed. The ten posts you could not tell apart were not written by lazy people or bad marketers. They were produced, correctly, by a system built to reward exactly what they delivered: volume, coverage, a checked box. I built part of the discipline those systems were supposed to scale. Marketing automation and agentic AI are forcing functions. They require a balance of human and machine judgment to scale effectively. Right now we are way, way out of balance, and the declining impact of content marketing proves it.

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

Uphoff on Media is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory, a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.

A golf equipment company just provided a realtime case study to everyone building, running, funding, or starting a B2B company. It’s the exact risk sitting inside every co-marketing deal, sponsored campaign, and reseller partnership on the books right now.

Callaway partnered with Good Good, the golf YouTube channel with roughly 4 million subscribers, to launch a new driver. The launch ad showed Good Good co-founder Garrett Clark running full speed into fellow Good Good staffer Alexis Miestowski, knocking her to the ground, then standing over her. It went out under both brands. It set off an immediate backlash, that prompted Callaway’s CEO Chip Brewer to make this public statement.

Read that sentence again. Not “a rogue post got out.” Not “someone published without sign-off.” Someone approved it. That’s the whole story, and it’s the same story sitting inside your partner marketing schedule right now, just without the headline.

Subscribe now

The stakes here weren't small. Equipment is roughly 70% of Callaway's revenue and the company's highest-margin segment: about 23% operating margin, worth roughly $100 million in operating income in a typical quarter. The market didn't wait to see how this played out. Callaway's stock dropped 2% the day the story broke, a one-day loss of about $57 million in market value, more than half of what the entire equipment division earns in a full quarter.

Dick’s and Golf Galaxy didn’t just pull the ad. They pulled every piece of Good Good merchandise from shelves and websites. The PGA Tour, set to have Good Good title-sponsor a new event this November, is now calling that sponsorship “a fluid situation.” Golf Channel postponed a show built around the partnership. None of that damage came from the driver. All of it came from one piece of co-branded content nobody stopped in time.

The partner had the pen. Callaway kept the liability

Good Good produced the ad and kept creative control. Callaway got access to Good Good’s platform and audience, which is the entire point of a deal like this: borrowed authenticity, borrowed reach. But Callaway still held approval rights and used them. The moment the ad published, Callaway owned 100% of the retail fallout, the sponsorship fallout, and the market reaction, despite writing none of the content.

That’s the actual structure of most co-marketing in B2B too. Sponsored content on a partner’s channel. A joint webinar where the guest controls the slides. A reseller’s campaign carrying your logo. A sponsored analyst report. You get named. You rarely get final cut. The exposure transfers with the logo. The control doesn’t.

Nobody wanted to slow down the growth

Good Good went from roughly 800 retail doors to 4,000 in the past year, and its merchandise was reportedly among the top sellers at Dick’s and Golf Galaxy in 2025. That’s the detail that actually explains the failure. Someone at Callaway had the authority to kill that ad before it published. Almost nobody in that position wants to be the one who slows down the fastest-growing partner in the building, especially when that partner is driving the growth number everyone in the room gets credit for.

That’s the more accurate version of “there was no process.” There usually is a process. What’s missing is a reviewer with the standing, and the incentive, to actually use it against a partner who’s driving the most growth for the business.

B2B runs on the identical structure, with less obvious violence

Partner-sourced revenue makes up roughly 24% of new bookings in horizontal SaaS, and considerably more in hardware and services-led businesses, according to a 2026 Crossbeam and Partnership Leaders survey of more than 2,100 vendors. Co-marketing isn’t a side channel anymore. It’s foundational.

Most of what goes wrong in that channel won’t look like a golf ad. It’ll look like an over-claimed stat in a partner’s co-branded blog post, a competitive claim in joint ad copy that draws a legal letter, a webinar co-host who goes off script on something the company never wanted attached to its name, a sponsored analyst report built on a number nobody at the sponsor actually verified. Same structural gap. Lower drama. Same exposure.

The speed of that content is changing too. Partner and creator content publishes in hours. Most brand review chains were built for campaigns that assumed weeks. I’ve written about the operating shift agentic AI is forcing on B2B workflows in “Loop Engineering”, and this is the same compression showing up in governance instead of production. AI didn’t create this gap. It’s about to widen it, by adding a third content source, alongside partners and creators, that moves even faster and has the same unclear owner of the veto.

The relationship rarely stays contained to the campaign

Retailers didn’t just pull an ad. They pulled a brand. The Tour didn’t distance itself from a video. It put a signed sponsorship in question. That’s the part most people running these partnerships underprice: a co-marketing misfire with a channel partner or reseller rarely stays bounded to that one campaign. It puts deal registration, MDF, roadmap access, and the whole commercial relationship in play, because the retailer or the platform doesn’t distinguish between the partner and the content once it’s gone wrong.

And that’s the visible version, the kind that makes headlines. Most B2B brand damage doesn’t. A prospect or customer who’s uneasy about a partner overstepping, or a claim in a co-branded campaign that felt off, isn’t going to call and tell you that. B2B buyers don’t behave like consumers: no public complaint, no organized boycott, no social media pile-on.

What happens instead is quieter and harder to trace. A deal that was moving stalls. A shortlist invitation that used to come automatically doesn’t. Nobody names the reason, because nobody has to, and leadership reads it as a soft quarter instead of the actual cause.

That’s what makes it harder to recover from than Callaway’s version. A public blowup at least tells you exactly what broke, in front of everyone, with a clear moment to respond. Quiet erosion of trust sends no signal at all. You lose the account before you know you had a problem.

Where the checks and balances actually have to live

Most companies don’t need a new marketing process. They need to know where in the existing one the veto actually sits, and what automatically escalates regardless of who produced the content. A simple stage-gate framework does that: concept and partner selection, production, pre-publish review, publish and distribute, monitor and escalate. At each stage, name who holds creative control and who owns the risk if it goes wrong. Those two answers are almost never the same name, and that gap is exactly where governance has to live.

The review stage is where it actually gets decided, and it only works if a fixed set of triggers pulls content to a senior sign-off automatically, regardless of who made it or how well the relationship is performing: anything depicting physical harm, anything touching discrimination or a protected class, any competitive or financial claim that hasn’t been verified, any conflict with an exclusivity or IP commitment. Take the judgment call out of the hands of whoever’s most invested in not slowing the partner down.

Download Who Holds the Pen, Who Holds the Risk →

Callaway didn’t lose control because nobody was watching. It lost control because the person watching had every reason not to slow down the fastest-growing thing on the shelf. That’s not a golf problem. It’s the exact shape of every co-marketing deal on your calendar, and the fix isn’t slower marketing. It’s making sure the person holding the veto has both the authority and the incentive to use it, especially against the partner everyone’s most excited about.

Subscribe now

The views expressed in Uphoff on Media are entirely my own. They don’t represent the opinions of any company I’ve led, any board I’ve sat on, or any investor who’s had the pleasure of debating strategy with me over the years. If something I write here sounds brilliant, I’ll take full credit. If it turns out to be wrong, I was clearly misquoted by myself.

”Uphoff on Media” is published by Tony Uphoff, Founder and Managing Partner of Uphoff Advisory a strategic advisory practice for founders, CEOs, and investors in B2B information, marketing, and technology. The businesses that drive business.