To niche or not to niche: How specialised should your business be?

"Pick a niche" is right but incomplete — being specific makes you easier to refer, trust and charge for, but the advice falls apart the second you stop at "pick one."

The obvious niches are usually the wrong ones — geography, sector and service line are the units people reach for, but the better choice is almost always a type of client or type of problem.

Niching backfires three ways — when it's too early (no data yet), when it's driven by anxiety (clients can feel a desperate choice), and when you pick the wrong unit.

Differentiation has three levels — category (set for you), specialisation (where you plant the flag), and qualifying characteristics (your secret sauce: proprietary process, IP or premium pricing). The real work is in level three.

Your niche needs a big enough pond — David C. Baker and Blair Enns put the sweet spot at roughly 2,000–10,000 potential clients, though it flexes with your model. Michelin-starred and McDonald's are both viable, but they need very different numbers.

Geography is a lens, not a niche — one agency's best clients were place-rooted, community-minded businesses that existed everywhere, not just their county. What you're great at travels; where you happen to sit doesn't.

A niche is a treatment, not a diagnosis — wrong leads are a messaging problem, good chats that don't convert a credibility problem, too few leads a visibility problem. Diagnose first, then decide what to niche on.

The answer is hiding in your best clients — the work that was good, the relationships that were easy, the projects you'd do again. Find what they share and you've probably found your niche.