Finding a profitable niche

A niche that "looks good" is recognisable by this: it is exciting, and nobody has a figure to defend it. A niche that pays holds up against five questions. Here they are, in the order that eliminates fastest.

These five criteria assume you already know where to find the data. If you do not, start with the six-step market research method: it covers what to measure, and in what order.

Criterion 1: Does demand clear your floor?

This is the question that eliminates the most ideas. There is no universal floor: it depends on your unit margin and on what a customer costs you. A €300 product lives comfortably on demand that the same model would find unworkable at €12.

What matters is setting the floor before looking at the numbers. Otherwise you will adjust it to fit the idea you already like.

For a sense of scale, MarketRaccoon flags any demand below 100 monthly searches in red, and refuses green below 300. That is our convention, not a rule of commerce, but it reflects a simple idea: below that level there is no market to capture, there is a market to create. That is a different job, and a different budget.

The opposite trap exists too. Very high volume is not a good sign in itself: "shoes" is searched enormously and is not a niche. It is a sector, and it is already taken.

Criterion 2: Is the trend rising or falling?

Compare demand over the last twelve months with the period before. A growing share of a shrinking market is still a bad deal: you run to stay in the same place.

Be careful not to mistake a decline for a season. An outdoor product measured in November always looks moribund. Look at the full twelve-month curve, and note where the peak sits relative to today, before concluding anything.

Criterion 3: Does the competition leave room?

The number of sellers tells you nothing. Three players holding most of the market and forty splitting it evenly are opposite situations, and in both cases you would have written "a lot of competition".

The useful measure is the top seller's share. MarketRaccoon calls a market "concentrated" beyond 30% of the collected offer. Above that, someone sets the reference price and your positioning is defined relative to them. Below it, there is room, but ask yourself why nobody has managed to pull ahead.

A measurement precaution: a seller appearing with a single listing is noise, not a competitor. We only track one from three products found onwards, otherwise they drift in and out of the data at random. The full reading (structure, real prices, weak points) is in our guide on analysing the competition.

Criterion 4: Is there margin left once everything is counted?

Take the market's median price, then do the full subtraction: selling price, minus cost of goods, minus shipping, minus platform fees, minus the cost of acquiring a customer.

That last item is the one systematically forgotten, and it is almost always the one that decides. A margin that looks comfortable on paper disappears entirely if acquiring the buyer costs more than they bring in, and volume does not help, it simply multiplies the loss.

Do not look for a magic percentage: work out your own, and check that it leaves enough to fund the next sale. A niche that does not generate enough to pay for its own growth is a job, not a business.

Criterion 5: Can your cash flow absorb the seasonality?

A niche can be excellent and unliveable if most sales land in a few weeks. That is not disqualifying, but it changes everything about cash flow: you buy stock before the peak and live the rest of the year on what that period brought in.

Look at where the peak sits relative to today. If it has just passed, the decline you are measuring is seasonal rather than structural, and you have almost a year to prepare for the next one, which is rather good news.

How to combine the five

A niche that clears all five without reservation is rare. In practice you are looking for an idea that clears four, whose fifth weakness you know how to offset.

• Low demand but high unit margin: workable, provided acquisition cost follows.
• Concentrated competition but a clear product angle: workable, you are not entering head-on.
• Strong seasonality but solid cash reserves: workable, it is a calendar problem.
• Falling trend: no. The other four do not make up for that one.

Why you must evaluate several

The real trap is not picking a bad niche: it is evaluating only one. Effort already invested in an idea pushes you to validate it, and the five criteria then become a formality you let it pass with indulgence.

The remedy fits in a sentence: put the five questions to five or ten ideas before going deep on one. The same data serves them all, and the comparison makes obvious what studying one in isolation made arguable.