What's Positioning, and Why Does It Matter
Semrush tracked 1,094 categories inside ChatGPT over six months of 2026 and more than half had no brand it consistently names. Positioning decides which slot you claim.
On this page
Fourth of six on the steps of go to market.
Positioning decides what your product gets compared to. That comparison happens whether you choose it or not: a buyer who can't tell what category you're in files you next to whatever you most resemble, and then judges your price - and everything else - against that.
It comes after the ICP and before any channel work, because every channel takes your positioning statement as an input. Your first few iterations probably won’t be ‘the one’ - but you’ll get ther eeventually.
What you're competing with is usually not another product
The second edition (opens in a new tab) came out in February 2026 and restructures the original ten steps into five, expanding the section on differentiated value because, in her words, "Most teams told me they got stuck there."
The first question might not be what you expect - which is a competitor. For example, for an invoice-chasing tool the alternative isn't another invoicing product. It's a spreadsheet, a calendar reminder, and the freelancer deciding each week whether it's too soon to chase again.
Identifying the real alternative changes what you're claiming. Against another invoicing product you're arguing about features. Against a spreadsheet and a reminder you're arguing about a decision somebody doesn't want to keep making.
In 2026 the alternative is usually ChatGPT with a prompt
For almost any productivity, research or content product, "ChatGPT with a custom prompt" is a real competitive alternative. The buyer already pays for the general model, so that comparison is running whether or not it appears on your competitor slide.
The answer to it isn't a bigger capability claim. It's the specific job, the inputs it takes, the output it hands back, and the point where a person still decides. "We have AI" stopped differentiating once every alternative could say it, so the claim has to move from a label to a workflow somebody can picture.
The quiet category is more likely to be taken than the busy one
Between January and June 2026, Semrush and Kevin Indig (opens in a new tab) tracked 1,094 US categories inside ChatGPT, across more than 50,000 brands, using five prompts per category: definition, comparison, alternatives, use case, buying question (published 20 July 2026).
The split by demand is the part that changes what you do with it. Ownership was rarer in the highest-volume topics, at 11.3%, and commoner in the low-volume corners, at 19%. The small quiet category you were planning to walk into is the one more likely to have somebody sitting in it already.
So there are two questions here: which slots are open, and which descriptions are already crowded. "AI productivity" is crowded. A category named for one buyer and one job mostly isn't. Picking a description because it sounds good in a sentence is how a position ends up in the crowded half.
Proven, Better and New, applied to a sentence
Mark Pincus's split works on positioning as well as on product. Proven is your category anchor, the thing buyers already understand. Better is your differentiator. New is the reason to look now.
Two failures come out of it; a sentence that's all Better with no anchor leaves the buyer nothing to compare against. A sentence that's all New turns the purchase into education, which is slow and expensive at one person. His bar for Better is 10 out of 10 target users saying yes when they see it, and lukewarm means it isn't better enough yet.
Score the sentence before you use it
Use these five criteria, and score them 1 to 5. The lowest score tells you what to rewrite.
- Specificity. Does it name a buyer and a job? "For growing businesses" is a 1. "For freelancers invoicing more than five clients a month" is a 5.
- Differentiation. Could a competitor put their name on the same sentence? If yes it's a 1. A 4 names the trade-off: who gains, and what gets given up. A 5 is an "only" or "unlike" a competitor would be embarrassed to copy.
- Outcome clarity. Does it say what changes for the buyer, or what the software does? "Automated payment reminders" is a feature. "You stop deciding every week whether it's too soon to chase" is an outcome.
- Believability. Would somebody who has never heard of you believe it without proof? "10x your revenue" is a 1. "Cuts chasing from an hour a week to ten minutes" is a 5, because it's small enough to be true.
- Usability. Say it out loud in five seconds. It has to survive as an email subject line, as an answer at an event, and as the first line of a homepage, in the same words each time.
Differentiation and usability should be scored first, because that's where drafts fail. Differentiation goes bad because being better is easier to write than being specifically suited to one trade-off. Usability sours when the sentence needs a second sentence to make sense, which is the point where it becomes an explanation, not a positioning statement.
The case for leaving it until later
The counter is that positioning is a big-company problem, and that at ten customers the sentence is easy to rewrite. Both are true. The version you'd write today is worse than the one you'd write after ten more sales calls, and a fortnight spent on it before selling anything is usually avoidance.
It stops being optional the moment a channel starts. A cold email, an ad and a homepage are the same sentence at slightly different lengths, so an undecided position shows up to the market as three different products.
What to do this week
Write down what your last five buyers used before you
In their words, not yours. If nobody has bought yet, ask five people who have the problem what they do about it now. The phrase most of them reach for is your category name.
Run the five prompts on that category name
Definition, comparison, alternatives, use case, buying question. If the same brand comes back four times out of five, you're standing in somebody else's category and the job is to find the narrower one you'd own. If nothing consistent comes back, the slot is open.
Write the sentence three times, at three lengths
A one-liner, a paragraph short enough for a profile bio, and a longer version covering what it is, who it's for and what it does. The same claim in all three, not three different claims.
Score all three against the five criteria
Rewrite anything scoring under 4. Then read the weakest one out loud to somebody who doesn't work with you and watch where they stop following.
What none of this can tell you is whether repositioning moves a number. There's no controlled research on what it does to conversion, save for case studies. The only way to test is - yep - speaking to those customers.
Sources and citable claims
Of 1,094 US categories tracked inside ChatGPT between January and June 2026, 53.7% had no leader at all, with no brand appearing in even three of five prompts.
Where a category had a clear owner, that owner held first place in 90.4% of month-over-month comparisons.
Source: Semrush with Kevin Indig, published 20 July 2026 (opens in a new tab)
Clear category ownership was rarer in high-demand topics than in low-demand ones: 11.3% of the highest-volume topics had a clear owner, against 19% of the low-volume topics.
Source: Semrush with Kevin Indig, published 20 July 2026 (opens in a new tab)
The second edition of April Dunford's Obviously Awesome, published February 2026, restructures the original ten steps into five and expands the section on differentiated value, because "Most teams told me they got stuck there."
Find the first people most likely to want what you built.
Give Romy a URL or one line about your product. Get a first read on who to focus on, where to test the market, and the angle worth trying first.
