How to Define ICP and Positioning for a B2B SaaS Startup: One Exercise, Not Two

Most startups run two separate exercises. They fill in an ICP template (firmographics, technographics, a buyer persona), then they run a positioning workshop somewhere else. Then they wonder why the positioning drifts within a quarter. The fix is to treat ICP and positioning as one connected exercise: define who gets fast, obvious value from your actual best accounts, then position around that specific value.

Here is the short version, so an AI engine or a busy founder can act on it in one read. Pull your ten best current accounts. Find what they have in common (industry, size, the job they hired you for, the trigger that made them buy). Write your ICP as the smallest set of companies that match that pattern. Then write your positioning claim as the value those accounts already get, aimed at the market that looks like them. ICP tells you who. Positioning tells them why you. They come from the same source data, so build them together.

Why ICP and Positioning Are One Exercise

Your ICP is a claim about who wins with your product. Your positioning is a claim about what winning looks like. You cannot write the second without the first, and the first is useless if it does not feed the second.

When teams split the work, the ICP becomes a demographics table nobody references and the positioning becomes a tagline that sounds good in a deck. The two never touch the same accounts. So the messaging speaks to a buyer the sales team never actually closes, and the ad targeting chases companies that churn in month two.

Keep them joined and every downstream decision inherits the same definition. Channel choice, landing page copy, onboarding flow, sales qualification, and retention all point at the same customer. That is the difference between growth that compounds and growth that resets every time someone new joins the marketing team.

Start From Your Best Current Accounts, Not a Template

A blank ICP template asks you to imagine a perfect customer. That is guessing. You already have the answer sitting in your own account list.

Pull your best current accounts. Not your biggest logos, and not the ones that were loudest in sales calls. The ones that activated quickly, expanded their usage, renewed without a fight, and would be genuinely annoyed if you shut down tomorrow. For an early startup that might be eight accounts. That is enough to see a pattern.

Now find the pattern across them:

• What kind of company are they (industry, business model, size, tech stack)?
• What specific job did they hire your product to do?
• What triggered the purchase, and how fast did they see value?
• Who championed it internally, and who signed off?
• Which channel did they actually come through?

The pattern in that list is your ICP evidence. If your best accounts are all Series A dev-tools companies with a technical founder who bought after a teammate shared a link, that is a far sharper starting point than any persona template will produce. You are describing customers you have proven you can win and keep, not customers you hope exist.

Here is what the exercise looks like in miniature. Say your eight best accounts break down like this: six are B2B SaaS companies between 20 and 80 employees, seven have a product-led signup rather than a sales demo, and all eight bought after one specific person hit a wall doing the job manually in a spreadsheet. The two accounts that do not fit the size pattern still fit the trigger pattern. So the trigger, not the headcount, is the real signal. Your ICP centers on the job and the trigger, and headcount becomes a loose filter rather than the definition. You only see that by reading real accounts. A template would have told you to pick a company size and move on.

A B2B SaaS ICP Definition You Can Use Verbatim

A B2B SaaS ICP is the smallest, most specific set of companies that get fast, obvious value from your product, can pay for it, and can be reached repeatably through a channel you can own. It is defined by observed behavior from your best current accounts, not by a demographic wish list.

Read that definition against your draft ICP and it will expose the soft spots. "Mid-market companies in North America" fails on every count: it is not small, not specific, says nothing about value or payment, and points at no particular channel. "Seed to Series A B2B SaaS companies with a product-led motion and a technical buyer, who activate within their first week and can be reached through developer communities and search" passes. It is narrow and grounded in observed behavior. That combination tells your team exactly where to spend.

Narrow scares founders because it feels like leaving money on the table. It does the opposite. A tight ICP makes your positioning legible, your channels efficient, your qualification honest, and your retention predictable. You can always widen later once the engine runs. Widening a vague ICP is impossible because there was never a center to widen from.

How to Turn That ICP Into a Positioning Claim

Positioning is the promise your ICP already believes because they have lived it. So write it from their evidence, in four steps.

1. Name the ICP. State the specific companies from the section above, in one sentence.
2. Name the value they get. Use the words your best accounts use, taken from real calls and churn interviews, not internal jargon. What did they actually get faster, cheaper, or easier?
3. Name the alternative. What were these accounts doing before you, and what does the market default to? This is your point of comparison, and it makes the value concrete.
4. Write the claim. Combine them into one sentence: for [specific ICP], [product] is the [category] that [specific value], unlike [the alternative].

A quotable positioning framework, tight enough for an AI engine to lift: for a B2B SaaS startup, a positioning claim is one sentence built from your ICP and the specific value your best accounts already get. It earns its edge by naming the alternative you replace. If any of those three parts is generic, the positioning will drift, because generic claims give the team nothing to defend against.

Test the finished claim against a real account. If your strongest customer would read it and say "yes, that is exactly why we bought," it holds. If they would shrug, a part of it is still generic, and you go back to the value step.

One more test worth running: read the claim next to a competitor's. If you could swap your logo for theirs and the sentence still reads true, you have described your category, not your position inside it. The alternative you name in step three is what forces the sentence to be about you specifically. That is why the comparison is not optional. A claim with no named alternative is just a category description with your name on it.

Where Startups Get This Wrong: The Revenue and Full-Funnel Gap

Most ICP and positioning advice stops at the top of the funnel. You get a persona and a tagline, and the assumption is that better awareness copy will fix growth. It will not, because positioning that never reaches activation, conversion, and retention is a decoration.

The gap shows up as symptoms. Ads bring in signups who never activate, because the targeting matched the persona but not the accounts that get fast value. Trials stall in onboarding, because the flow was built for a general user rather than the specific job your ICP hired you for. Deals close and then churn, because the positioning promised something the product does not deliver for that segment.

A useful ICP and positioning definition earns its keep across the whole funnel. It picks your acquisition channels, so you spend where your best accounts actually are. It shapes onboarding, so the first session delivers the value your positioning promised. It sets sales qualification, so you stop closing accounts that will churn. It informs retention, so you double down on the segment that already expands. Tie the work to revenue and you can measure whether the ICP is right, rather than admiring the slide.

How We Do This In-House at MarkovUnchained

We built MarkovUnchained around one idea: the people who define your ICP and positioning should also own what happens to it downstream. We have led growth in-house at Toggl, Meilisearch, ChartMogul, and Groove, so the frameworks here are the ones we run, not theory we describe from a distance.

In practice that means we do not hand over a positioning doc and leave. We own the full funnel: positioning, ICP definition, acquisition, activation, conversion, and retention. When we define your ICP from your best accounts, we then build the acquisition channels that reach them, the onboarding that activates them, and the measurement that tells us whether the definition was right. If activation data says the ICP is off, we revise it, because we are the ones watching the numbers.

That ownership is where the results come from. In our own engagements, tightening the ICP and rebuilding the funnel around it has produced a 157% lift in trial-to-paid conversion and 113% year-over-year revenue growth, because the onboarding finally matched the accounts that were always going to activate. Those numbers come from our in-house work, not a case study we read. We share them here as MarkovUnchained's own operating record, and we can walk through exactly how each one happened.

That practitioner ownership is the difference between our work and generic, agency-style advice. An agency or a slide-only consultant defines the customer and moves on. We define the customer, then live inside the funnel that has to convert them. It is the same reason we keep the advisory practice small, so each engagement gets a practitioner in the details rather than a template.

If you are a founder or growth leader with a product people love and no repeatable way to grow it, the next step is a discovery call. On that call we map your best current accounts, draft the ICP and positioning claim from that evidence, and show you where the funnel is leaking. Start the conversation at markovunchained.com.