What a niche market actually is
A niche market is a group of buyers with a shared requirement that the general-purpose products in their category treat as an edge case. The requirement is not exotic to them — it is the centre of their working day — and it is peripheral only from the point of view of a product built for someone else.
The distinction from a small market matters because the two behave in opposite ways. A small market is small for everyone, including you. A niche is small only for the incumbent: it can be the whole of your business while being a rounding error in theirs, and that asymmetry is the entire strategy.
Four properties of a profitable niche
A niche worth entering has all four. Missing any one of them produces a recognisable kind of failure, and the failures are more instructive than the successes.
- The buyers can find each other. A niche whose members have a forum, a trade association or a hashtag can be reached for the price of showing up; one with no gathering place has a customer acquisition cost that eats the margin.
- The requirement is expensive to bolt on. If the incumbent could satisfy it with a settings toggle, they will the moment you prove the demand. The defensible niches need a different data model, a different workflow, or a compliance burden the incumbent does not want.
- Money already moves through it. The members are paying for something today — a general product they tolerate, a consultant, or an assistant doing it by hand. A niche where the current solution is unpaid effort is a niche with no budget line.
- It has an edge you can grow along. The best niches are the narrow end of something wider, so that solving the hard case first earns the right to the easier ones later rather than trapping you at the size you started.
How to find one from public evidence
Niches announce themselves in the reviews of general-purpose products, in a specific grammatical form: a compliment followed by a disqualification. Great tool, but it cannot handle our case. That sentence is a segment describing the exact boundary at which a product stops working for them, written by the person on the wrong side of it.
Collect those sentences across every serious product in a category and sort them by the reason given, not by the product. Reasons that recur across all the products in the category are niches; reasons that appear against one product only are that product's bugs, and building a company on a competitor's bug list is a plan with a six-month shelf life.
Language and geography are the most reliable niches
The most durable underserved segment is usually a place, because localisation is exactly the kind of cost an incumbent postpones indefinitely: it needs local payment methods, local invoicing rules, local support hours and a translation that a native speaker does not wince at, and each of those is real work for a fraction of the incumbent's revenue.
This is why the Vietnamese segment of an otherwise global category is so often a real opportunity rather than a wish: the demand is already documented in Vietnamese, in reviews and forum threads that English-language research never reads, and the incumbent has read them even less than you have.
The test for a market that is simply too small
Take the number of distinct people you can actually count in the evidence, not an estimate. Multiply by the annual price the segment already pays for its current workaround. If that figure does not cover the cost of building and supporting the thing for a year, the niche is a hobby, and no amount of enthusiasm changes the arithmetic.
The arithmetic is deliberately crude. Its job is not to forecast revenue, it is to catch the case where the answer is off by a factor of ten, which is the case that actually costs people a year of their life.