How to niche the right way

#The formula, and why it's structured this way specifically

The standard practitioner formula for a checkable positioning claim is "I help [specific type of client] [achieve a specific outcome], usually by [your method]." Both halves matter, and each is doing distinct mechanical work described in why generalist positioning fails:

• The client-type half (Y) narrows who can vouch for you. "Series A SaaS companies" is a small, densely-connected world where people talk to each other — a referral inside it is worth more than a referral from a stranger, because the referrer and the prospect likely share context. "Businesses" is not a world; it's everyone, which means a referral inside it carries almost no shared context and is barely more informative than a cold lead.
• The outcome half (X) narrows what a track record can prove. "Cut CAC by rebuilding onboarding" is a claim three past clients can confirm or deny in a specific, checkable way. "Grow the business" cannot be confirmed or denied by anyone, including you.

Framework popularizers in this space (Blair Enns' "Win Without Pitching" positioning work, David C. Baker's writing on specialization) converge on materially the same formula, though neither publishes a controlled comparison proving the format itself outperforms alternatives — treat the formula as a well-tested practitioner convention, sourced in why it should work from the credence-goods mechanism above, rather than as an experimentally validated claim. [Directional — convergent practitioner convention, mechanistically grounded, not independently tested]

#How to actually pick the Y and the X, without guessing

Three inputs, in order of reliability, for choosing what to niche into:

1. Where you already have a checkable track record, even an informal one. If you've solved a version of this problem for two or three organizations already — as an employee, a volunteer, or a side project — you already have the referral-and-case-study raw material the mechanism above depends on. This is the strongest starting point precisely because it doesn't require you to manufacture credibility from nothing.
2. Where the client pool is dense enough to have referral chains at all. A niche needs enough people who know each other — a professional association, a founder community, an industry with regular conferences — for the referral mechanism to actually fire. A technically narrow niche with no social density (say, "the three companies in the world that do X") has the checkability property but not the referral-volume property, and needs a different first-client strategy — see The first-client problem.
3. Where the client's need is expensive enough to justify hiring outside help at all — recall condition 1 and 2 from the mechanism: the need has to be either intermittent or genuinely novel to the buyer for outside help to beat in-house effort on cost. A niche where the problem is cheap and constant is usually one the client eventually builds in-house rather than keeps buying externally.

#Niching is a starting claim, not a life sentence

The common objection — "what if I pick wrong, or the niche dries up" — is real, and the honest answer is that a niche is a starting hypothesis you test with real client conversations, not a permanent identity. Independent consultants routinely narrow into one niche to build the first checkable track record, then either expand the claim once credibility exists (the reused-playbook effect AI Agency's three-paths lesson documents for its own domain — the second-through-Nth client in the same niche is faster to scope and deliver than the first, which is exactly the capital a narrow start builds) or deliberately pivot to an adjacent niche once the current one is saturated or no longer interesting. The cost of picking a niche that turns out wrong is a few months of a specific positioning statement — not, as the fear usually implies, a permanently foreclosed broader practice.