Positioning frameworks are decision structures, not data
April Dunford's framework is the one most B2B teams reach for, and deservedly. Five components: what customers would use if you did not exist, what you have that those alternatives do not, the value that produces, who cares most about it, and the market category that makes all of it obvious. Positioning as context-setting.
It is a good framework. It is also, strictly, a set of questions — and two of the five are empirical claims about people who do not work at your company. What would customers do without you. Which customers care most. Those have answers that exist in the world and can be checked.
In practice they get answered in a room. Founders, a couple of marketers, maybe sales leadership, a whiteboard, an afternoon. Everyone in that room has already crossed the curse-of-knowledge threshold, which is precisely the disqualifying credential.
There is a deeper problem with the room, though, and it is not the people in it. A messaging workshop optimises for internal agreement. Customer interviews optimise for external recognition. Those are different objectives, and only one of them is the objective. A workshop can produce a sentence everyone in the company genuinely likes while removing, by consensus, the exact words a buyer would have used to recognise themselves — because distinctive language is the first thing a room sands off on its way to something nobody objects to.
Untranslated, not wrong
What comes out of that room is usually accurate. The category is plausible, the differentiation is real, the value statement is defensible. It describes the product correctly from the vantage point of people who built it.
The buyer's vantage point is different. They are trying to make progress before they are trying to classify vendors. They are not evaluating a category; they are trying to get something done, and they have their own words for it — words formed before they knew your product existed, in conversations with colleagues, in the way they describe their week to their manager. When your messaging arrives in different vocabulary, nothing signals to them that it is about their situation. Not disagreement. Non-recognition.
This is the curse of knowledge doing its ordinary work. Once you understand your product deeply you cannot reconstruct what it was like not to, and the language that feels clearest to you is the language that assumes the most.
Elizabeth Newton demonstrated the mechanism in a study that has become the standard illustration. One group tapped out the rhythm of a well-known song; another tried to name it. The tappers, hearing the melody in their heads, wildly overestimated how often listeners would get it. The listeners heard tapping. That is the relationship between your messaging and a buyer encountering it for the first time.
The same job, four vocabularies
It compounds in B2B, because there is no single buyer. A purchase involves an economic buyer, the people who will use it, someone technical, and usually a champion carrying it internally. They are pursuing the same underlying job and they do not describe it the same way at all.
Take a product that catches errors in customer data before they reach reporting.
| Role | How they describe the same job | Actually optimising for |
|---|---|---|
| Economic | "We keep making decisions on numbers that turn out to be wrong." | Business outcomes |
| User | "I spend the first two days of every month cleaning the export." | Their own time |
| Technical | "There's no validation layer between ingest and the warehouse." | Risk |
| Champion | "I'm tired of being the person who explains why the dashboard is wrong." | Reputation |
Notice which one is missing from most messaging. Companies write for the user, because the user is who they talk to — support tickets, onboarding calls, feature requests. The economic buyer signs, the champion does the internal selling, and neither hears their own sentence anywhere on the site.
Notice also that the champion's version is not about data at all. It is about not being the person who has to apologise.
Champions are optimising for something the other three are not: status, credibility, and career risk. Backing a tool that fails is a professional cost they carry personally, long after the invoice is forgotten. That is the emotional and social side of the job, it is rarely written down anywhere, and it is the version most likely to get repeated in a meeting you are not in — which makes it the version that decides whether the deal survives internal scrutiny.
But customers can't tell you why they bought
This is the honest objection, and it is largely true. Ask someone why they chose a product and you will get a post-hoc account assembled to sound reasonable. Buyers are unreliable narrators of their own motives.
But that is a different question from the one that matters here. You are not asking them to introspect. You are collecting vocabulary — how they describe the situation, what they call the problem, which words they reach for when explaining it to someone in the same seat. People are poor witnesses to their reasoning and excellent witnesses to their own phrasing.
Which is why the single most useful question in this kind of interview is not why they bought. It is how they would describe this to someone in their role at another company. That answer is not analysis. It is the copy.
The diagnostic, and what to do about it
Open your homepage. Point at any sentence. Ask who said it first.
The candidate answers are short: a founder, a product manager, someone in marketing, someone in sales, or a customer. Four of those five are the same answer wearing different job titles. If you can name a customer and roughly when they said it, that sentence is doing work. Most B2B homepages cannot produce a single one.
The fix is not another positioning exercise, because the exercise is not what failed. It is a round of core jobs interviews across all four roles, run with a bias toward capturing exact phrasing rather than summarising it. Ten to twelve conversations, spread across economic, user, technical and champion seats, transcribed and read for language rather than for themes.
Read for frequency as well as phrasing. Which nouns recur across transcripts, which verbs people reach for, which metaphor keeps reappearing without prompting. A phrase three unrelated buyers used independently is worth more than an elegant one you constructed.
What comes back is a vocabulary list you did not write, sorted by who says what. Where the four roles diverge is where your messaging needs to fork. Where they converge is your headline, and it will usually be a sentence nobody in your company would have written unprompted.
The complete core jobs interview guide
Twenty-five questions across four phases, built to be run across all four buyer roles, with what each one is designed to surface. Free, no email required.
Open the guide →The five-component positioning framework and the description of positioning as context-setting are April Dunford's, from Obviously Awesome (2019). The curse of knowledge is a well-documented cognitive bias; the tapper-and-listener study is Elizabeth Newton’s (1990), popularised by Chip and Dan Heath in Made to Stick (2007). The argument that most positioning failures are language failures, the four-vocabulary framing, and the homepage diagnostic are ours.