Most founders treat positioning as a writing task. They draft a sentence, put it on a slide, and move on. Then demos stall, homepage copy feels generic, and AI tools describe the company in someone else’s language.
That sequence is common, and avoidable. Positioning is not a slogan. It is a falsifiable hypothesis about how a specific buyer should understand your product, why they should care, and what category they should place you in.
This matters most before product market fit, when evidence is thin and judgment is still expensive. The goal is not perfection. The goal is a one week test loop that gives you enough signal to sharpen the claim.
Why statements fail
Most startup positioning statements are written before the company has enough customer evidence. They sound coherent in isolation, but they do not survive contact with actual buyers.
Three failures show up repeatedly.
- The statement is too broad, so it fits no one.
- The language is founder language, not buyer language.
- The category is chosen for pride, not comprehension.
April Dunford’s work is useful here because it separates positioning from copywriting. Her core point is simple, positioning is the set of context and comparisons that make your product make sense, not the sentence itself, see her template on startup positioning.
Al Ries and Jack Trout made a similar case years earlier, category perception drives what people remember, and memory drives choice. The company that is first to occupy a clear mental slot gets an advantage. That does not mean the slot is fixed forever, it means the slot is real.
So if the statement has not been tested against buyer language, it is fiction with nice typography.
What positioning is
In practice, positioning answers four questions.
- Who is this for?
- What problem is urgent enough to buy?
- What frame makes the product intelligible?
- Why choose this over the obvious alternative?
A useful shorthand is this, positioning is the story a buyer repeats back when they finally get it.
Dunford’s canvas is still one of the best operating models for founders because it forces context. It asks you to define the customer, the alternatives, the market category, the differentiators, and the value. But even a good canvas becomes dangerous if you stop at the document.
The point is not to produce a polished artifact. The point is to identify a claim you can test.
Make it falsifiable
Use this format.
For [narrow ICP] who is stuck with [current alternative], we are [category] that helps them achieve [specific outcome]. Unlike [alternative], we provide [credible proof].
This is not meant to be poetic. It is meant to be testable.
Good hypotheses have boundaries. They force tradeoffs. They help you decide which words belong on the homepage, which audience deserves sales effort, and which claims can be defended in a call.
For example, a weak version sounds like this.
For modern teams, we are an AI platform that helps companies work smarter.
That sentence is not wrong. It is unusable.
A testable version sounds more like this.
For security-minded SaaS teams that still manage compliance manually, we are a workflow layer that turns audit prep into a recurring process, unlike spreadsheets and one-off consultants, we keep evidence current.
Now you can test the buyer, the alternative, and the promise.
Run the loop
You do not need a research department. You need a disciplined week.
Test one
Repeat-back on five calls
On sales calls, founder interviews, or customer discovery sessions, explain your positioning in one sentence, then stop. Ask one question, “How would you describe this to a teammate?”
You are not looking for praise. You are listening for exact language.
Pass conditions include:
- They restate the problem in their own words.
- They mention the right alternative without prompting.
- They can name the value in plain language.
Fail conditions include:
- They summarize you as a feature, not a category.
- They ask, “So what do you actually do?”
- They repeat your phrase without understanding it.
If you hear the same confusion three times, the hypothesis is weak, even if the sentence reads well.
Test two
Problem-led versus category-led
Rewrite the homepage hero in two versions.
Version one leads with the pain. Version two leads with the category.
You do not need traffic to learn something. Use direct messages, founder communities, customer emails, LinkedIn posts, or small paid experiments if you already have them. The question is not which headline is cleverer. The question is which one gets cleaner replies.
Problem-led copy often works early because buyers recognize their own frustration before they recognize a category. Category-led copy can work later, when the market already knows the frame.
The lesson is practical, not ideological. If the problem version creates more “yes, that’s us” responses, keep it. If the category version creates faster comprehension, keep that.
Test three
Mine lost deals
Look for three phrases.
- “We went another direction.”
- “We thought you were more like…”
- “I’m not sure what bucket this fits in.”
These are positioning clues, not objections to treat casually. They tell you whether buyers are placing you in the right mental shelf.
Also mine the softer moments. The comment after a demo, the question at a meetup, the introduction from a friend, the moment someone says, “Wait, what do you do again?”
These moments matter because they are unguarded. They reveal the mental model you have actually created.
Choose deliberately
Many founders think category choice is a branding decision. It is not. It is an operating decision.
April Dunford’s database to data warehouse story is useful because it shows how the same product can be understood through different frames, and how the frame changes perception of value. A product can be technically excellent and still be placed in the wrong mental bucket.
That means your category should do three things.
- Make the buyer feel immediate relevance.
- Make the alternative obvious.
- Make your proof legible.
Choose the narrowest category that still captures your real value. Early founders often try to sound expansive. That usually weakens comprehension. If your product solves a sharp problem for a sharp group, say so.
There is a reason broad labels like “platform” and “AI for teams” feel safe. They avoid exclusion. But exclusion is often what gives positioning force. If nobody is left out, nobody is clearly in.
The AI layer
There is a new reason to care about this discipline. Buyers are increasingly asking AI tools what your company is, what category you belong to, and whether you are credible.
In other words, the category you claim now shapes the category AI repeats later.
If your positioning is vague, LLMs tend to collapse you into generic language. If your positioning is crisp, specific, and repeated across your site, posts, product pages, and third-party mentions, you are more likely to be described back correctly.
This is one reason the old copy only approach is insufficient. A homepage can sound good to a human and still be ambiguous to an answer engine. The claim needs distribution, not just polish.
Test it directly. Ask ChatGPT, Perplexity, or a similar tool, “What does this company do, who is it for, and what category is it in?” Then compare the answer to your intent.
If the system calls you something too broad, too technical, or simply wrong, treat that as evidence. The market is reading you incorrectly, and the machine is echoing it.
This is also where published work helps. Better Marketing exists to help early stage software companies earn trust through published work, but the work only compounds if the market can classify you cleanly. Positioning is the claim the content has to keep proving.
When to move
Positioning should not change every week. It should change when the evidence changes.
Watch for these signals.
- Your best prospects all describe the problem differently from your copy.
- Your strongest use case is not the one you lead with.
- The buyers who convert most quickly are not the audience you targeted first.
- Your category label creates more explanation than recognition.
When those signals cluster, revisit the hypothesis. Not the whole company, just the framing.
A useful cadence is monthly for new founders, quarterly once you have repeat patterns. The rule is simple, leave the core promise stable long enough to learn, but do not confuse inertia with conviction.
What should stay fixed is the customer truth. What should move is the language around it.
Template and example
Use this template as a starting point.
For [specific ICP] who currently rely on [current alternative], we are [category] that helps them [job to be done]. Unlike [main alternative], we [proof or mechanism].
Now a fictional example for a devtool company.
For small platform teams shipping internal tools without a dedicated DevOps function, we are a deployment control layer that reduces release risk before production, unlike manual scripts and generic CI, we keep approvals, checks, and rollback paths in one place.
That is not the final version. It is a working hypothesis.
Now test it against the week loop.
- Say it on five calls.
- Compare problem-led and category-led homepage versions.
- Mine the exact words from lost deals and casual confusion.
If the same phrasing keeps surviving those tests, you are getting closer to positioning that behaves like a market claim, not a writing exercise.
People ask
What are the 4 types of positioning?
There is no universal four-part list, but common dimensions include category positioning, problem positioning, audience positioning, and competitor positioning. Founders should use whichever lens clarifies buyer choice, not whichever list sounds tidy.
What are the positions in a startup?
If you mean organizational roles, startups usually compress responsibilities across product, engineering, sales, marketing, and operations. If you mean market positions, the company can occupy a category, a use case, a segment, or a niche.
What are the 4 pillars of positioning?
Many frameworks converge on four pillars, customer, problem, category, and differentiation. Those pillars matter only if they map to buyer language and a real alternative.
What are the 4 P’s of startup?
People usually mean product, price, place, and promotion. That model is useful for marketing execution, but it is not a positioning framework. Positioning comes before the tactical mix because it shapes what the product means in the first place.
Takeaway
Do not write a positioning statement and call it done. Treat startup positioning as a hypothesis, then test it against buyer language, homepage responses, lost deals, and AI summaries.
If the market cannot repeat your claim accurately, the claim is not ready.
Start narrow. Choose a clear alternative. Say the job to be done in plain words. Then run the one week loop before you polish the line.
That is enough to move from fiction to evidence.
First Round Review, Stripe, and April Dunford all point to the importance of clarity. The missing step, especially for pre PMF founders, is the test loop. Build that first, and your positioning becomes something the market can answer back to.