Who writes the criteria B2B buyers evaluate you on?

The scene I'm in this week

Tuesday morning, mid-August. A $34M B2B software company, PE-backed, selling quality management software to medical device manufacturers. The CEO forwarded me a spreadsheet at 7am with one line above it. How are we losing on this?

The spreadsheet was a vendor evaluation matrix a prospect's team had sent over. Forty rows. Weighted scores. Three vendors across the top. He'd scored third.

We went through it together. He won most of the rows he cared about. He lost row twelve and row nineteen, both about a workflow his team had deliberately removed two years earlier because customers kept misusing it and failing audits. He lost row thirty-one, a reporting module that exists at exactly one competitor and nowhere else in the category.

I asked him one question. Who wrote this?

He didn't know. The prospect didn't know either, as it turned out. Somebody on their side had assembled it from a few sources over a couple of days. I spent twenty minutes searching exact phrases from the rows. Eleven of the forty came back near-verbatim from a competitor's buyer's guide published in 2024. Six more traced to an analyst grid. The rest were internal.

He'd been arguing with a document his competitor wrote, in a language his competitor chose, and he'd been doing it for two years without knowing it.

Naming what's actually broken

The villain here has a name. I call it the Inherited Checklist.

The Inherited Checklist is the list of evaluation criteria a buyer walks into your first call already holding. Nobody on your team wrote it. Most of the time nobody on the buyer's team wrote it either. It descends from whoever in your category published a buyer's guide, a comparison page, or an RFP template, and it has been copied forward ever since, losing its author a little more with every pass.

That original publisher is usually a competitor with a content team and a stake in what lands on the list.

This is just truth: a checklist is never neutral. Every row on it is a decision about what matters, and whoever wrote the rows decided in their own favor. They put their strengths on the list and left their weaknesses off it. Then they gave it away for free, and the category adopted it as the objective standard.

This is Solution-Centric Marketing coming back around on you. When every vendor in a category describes itself by capability, the only comparison anyone can build out of that material is a capability grid. You handed the buyer a feature list. They built a scorecard out of it. Now you're getting graded on the exact thing you taught them to grade.

Naming the problem your company solves is the first act of authorship, and I wrote about that in Naming Rights: why the company that names the problem owns the market. The second act is naming the criteria. Most companies do neither, then wonder why the conversation never happens on their terms.

Answering somebody else's checklist well is close to the definition of being an option, because the list was built to compare options. Leading a rebellion in your category means publishing criteria that only make sense if your read of the problem is right, and letting the buyer decide whether they agree with you.

Why this is worse now than ever

Publishing a buyer's guide used to be a real investment. A content team, a quarter of runway, a designer, a gated PDF, a form. That cost worked as a filter. Only a few companies in each category bothered, so only a few checklists existed and you could usually tell whose was whose.

That cost is gone. A vendor can produce a plausible 2026 buyer's guide, a comparison page, and an RFP template in one afternoon. Every category now carries dozens of them, all confident, all slightly different, every one of them written by somebody with a stake in the outcome.

Here's what changed for you specifically. The buyer doesn't read those guides anymore. They ask the machine. And the machine doesn't hand back one competitor's opinion with a logo on it, which a buyer would discount on sight. It reads the whole pile and returns a single merged list in a calm neutral voice, sources folded in, no author left to argue with.

That merged list becomes the buyer's questions. It reaches your call as curiosity rather than as a document. Nobody emails it to you. You just notice, deal after deal, that prospects keep asking about the same three things and never once ask about the thing you're best at.

“In 2026, building is easy and distribution is hard ... the moat is execution: knowing your market, owning a customer relationship, building a brand that attracts the right users.”

... SeedScope, Vibe Coding Is How Startups Are Being Built in 2026

Owning the customer relationship starts earlier than most founders think. It starts the moment a buyer decides what a good vendor looks like, and that moment now happens weeks before anybody fills out your form.

The diagnostic: run this on your last five deals

Three tests. All of them run this week, with no budget and nobody hired.

1. 1The Source Trace. Take the most recent evaluation matrix, RFP, or vendor questionnaire a prospect sent you. Pick five rows and search the exact phrasing in quotes. If those rows land on a competitor's buyer's guide, an analyst grid, or a template blog post, you now know who wrote your test. Most founders find at least a quarter of the rows have a findable source, and it's rarely them.
2. 2The Default List. Open a fresh AI session with no account history. Ask it how a buyer should evaluate vendors in your category, and print the list it gives you. Read it twice. Count the rows you'd win on. Then count the rows that only exist because one competitor built one feature. That printed list is roughly what your next five prospects are walking in holding.
3. 3The Won-Deal Read. Call your last three closed-won buyers and ask what actually made them pick you, in their words, and write the answers down without cleaning them up. Now look for those answers on the checklist. When the reason a buyer chose you appears nowhere on the document they used to choose you, the gap is authorship, and no roadmap item closes it.

What I see across 100+ B2B companies

The pattern is consistent enough that I look for it now. The checklist decides who makes the shortlist. The deal itself gets decided somewhere the checklist can't see.

A checklist filters on what's easy to write down: features, integrations, certifications, seat counts, deployment models. What closes the deal is almost always a judgment about whether you understand this buyer's situation better than the alternatives do. That judgment gets made in a live conversation and it never has a row.

Which leaves most companies optimizing hard against a document that only controls the first cut, and improvising against the thing that decides the outcome.

The gap widens once you look at where the buyer's frame is being assembled from. Industry GEO research through May 2026 found that only 12 percent of top-ranking Google pages were cited by ChatGPT or Claude when asked the same questions. The pages that built your search presence are largely not the pages building your buyer's checklist. Two different libraries, and most companies have only ever stocked one of them.

The second pattern is the hopeful one. The missing criteria almost always exist inside the company already. A founder has said them out loud on a call. Something like: our customers are the ones who already failed an inspection once, so we design for the inspection instead of the workflow. That sentence is a criterion in raw form. It has just never been written anywhere a buyer or a machine could find it.

This sits next to two posts worth reading beside it. Why do buyers keep comparing us to the wrong competitors? is about who you get compared against. This one is about what you get scored on. And Why do competitors with weaker products win more deals than us? is what all of it looks like from the outside, when the weaker product wrote the exam.

A real example

PE-backed B2B SaaS, $34M in revenue, quality management software for medical device manufacturers. Roughly sixty percent of their deals ran through a formal evaluation matrix at some stage.

The Source Trace on four recent matrices turned up the same eleven rows every time, all traceable to one competitor's buyer's guide. The Default List, run cold in three separate AI sessions, put integration breadth first and audit readiness sixth in all three.

The Won-Deal Read is where it broke open. All three won buyers said a version of the same thing, and not one of them mentioned integrations. They picked this company because it had sat through FDA inspections alongside customers and had built the software around what an inspector actually asks for. That was the criterion. It had never appeared on a single matrix.

We ran the Magnetic Messaging Framework work first, because you can't publish criteria you haven't decided on. Villain framing gave them something specific to be against: quality software designed for the workflow instead of for the inspection. Category design gave a name to the thing they actually sell. Then they published their own evaluation piece, seven criteria, each written so it only holds up if their read of the problem is right.

Twelve weeks later, two prospects showed up at a first call asking about inspection design in their own words. Their sales lead told me the more useful change was quieter than that. Reps stopped defending rows they'd never chosen and started asking prospects why those rows were on the list at all.

What this means for you

Take one thing from this. The checklist your buyers use is a published document, which means it can be republished. Somebody wrote the current one because they decided to write it. You get to decide the same way.

The hard part sits upstream of the writing. You can't publish an evaluation framework you haven't settled internally, or you'll produce a seven-row list that's your feature roadmap with better verbs on it.

That settling is what the Magnetic Messaging Framework is for. It's a strategic narrative system built around four anchors: category design, villain framing, an old-way / new-way contrast, and a promised-land outcome. Two of those anchors carry the weight here. Villain framing settles what you're against, which is where a real criterion comes from, because a criterion is a belief about what goes wrong when somebody gets this decision wrong. Category design settles what you're comparable to, which decides which rows belong on the list at all. Once both are documented, publishing your criteria becomes a writing job instead of an argument.

Here's why that matters on a practical Tuesday. One of your reps is going to be handed a matrix next week. If the criteria live in a document, the rep can point at it and reframe the conversation in thirty seconds. If the criteria live in the founder's head, the rep defends rows one at a time and loses on volume.

I'm Greg Rosner, founder of PitchKitchen and author of Story Craft for Disruptors. We build Magnetic Messaging Frameworks for founder-led B2B companies in the $5M-$75M range, and we fix broken marketing messages and underperforming websites for CEOs whose sales are stalling because their message isn't doing the work. Most of what I actually do is pull a buried truth out of a founder's head and get it onto a page where a buyer and a machine can both find it.

Three things you can do this week.

1. 1Run the Source Trace on the last matrix you received. Twenty minutes and one browser tab. Find out whose document you've been arguing with.
2. 2Call three won buyers and write down the exact sentence each one uses for why they picked you. Leave the grammar alone. That sentence is your missing criterion in raw form.
3. 3Write seven criteria, not forty. Each one should be a sentence your closest competitor would hesitate to put on their own site. If they could publish your list unchanged, you've written a category checklist and you're back where you started.