Skip to content
WhittleOSWhittleOS
← All guidesProblems5 min read

Is my startup idea already taken? The wrong question, and the one that works

Half the results reassure you that competition is validation and the other half sell you a tool. Neither gives you a method, because yes-or-no cannot be acted on. Here is the question that can, and the 3 decisions it produces.

By Boris Binyaminov · ·

The wrong question
Has anyone built this
The answerable one
What do their users complain about
Possible answers
3, and none is yes or no
Unsourced complaints
Labelled, never dropped

Yes. Almost certainly, and it does not matter, because the question cannot be acted on either way. A no tells you nobody wanted it enough to try; a yes tells you nothing about whether there is room. The answerable question is narrower: are the incumbent's own users complaining about one specific thing — and the honest output is a decision among 3, not a yes or a no.

Why "taken" cannot be answered usefully

Half the results for this question reassure you that competition is validation. The other half sell you a tool that counts competitors. Both are answering a yes-or-no, and neither answer changes what you do on Monday.

The reassurance is not even wrong, it is just empty: competition is evidence that somebody believed in the market, which is a much weaker claim than it sounds when you notice how many crowded markets are crowded with unprofitable companies. And the count is worse than empty. Twelve competitors in a market you can reach beats two in a market you cannot, and no tally distinguishes those.

What you actually need to know is whether there is a specific, repeated, unfixed complaint — and whether it is one the incumbent will not fix, rather than one they simply have not got to yet.

The three answers, and what each one is

compete
Take them on where they are
The complaints are about something they will not fix, and you can reach the same buyer.
niche_down
Serve a slice they are not built for
The complaints come from one kind of user the incumbent is too general to serve well.
avoid
Leave it
The complaints are about the category itself, or nothing specific comes back at all.

The three-way verdict our teardown emits, rendered from its own vocabulary. None of them is "the idea is taken" — that is not a decision anyone can act on.

Notice what the middle one costs you. Narrowing to the slice an incumbent serves badly is the answer most often available to one person, and it means giving up the market you were imagining for the market you can actually win. That is a real loss, and a tool that never delivers it is flattering you.

Notice also that "leave it" is on the list, and that the trigger for it is often not a strong competitor. It is silence: you went looking for complaints and found none. That usually means the problem is not painful enough to write about, which is a finding about the market rather than about the incumbent.

One teardown, end to end

Take an appointment-reminder product for independent clinics. An incumbent already sends reminders, so “taken” is trivially yes. The useful work starts after that answer:

Complaint to verify: small clinics say the incumbent's setup assumes a multi-location administrator.

Evidence bar: the same workflow complaint on separate pages, from people already paying for the category.

Decision: niche_down only if that evidence exists and the incumbent is structurally unlikely to serve the slice; otherwise avoid.

First move: test the single-clinic workflow, not a second general reminder platform.

An illustration of how to reach a verdict, not evidence that clinics have this complaint. The missing source pages are exactly what the teardown must find before it can keep the idea.

The part that decides whether any of this is worth anything

You can produce all three verdicts from imagination, and most competitor analysis is exactly that — plausible weaknesses attributed to a company nobody checked. So the rule that matters is not about the verdict, it is about the evidence under it.

1
from a real page
Read off a page our fetcher opened in full.
2
you pasted this
Read off text you pasted in.
3
not verified
Inferred. No source. Marked so it can never read as a finding.

Every complaint carries one of these, in the same words the rest of the site uses. The third is the important one: it exists so an inference cannot be mistaken for a finding later.

The teardown's first non-negotiable rule is that it invents nothing. Concretely, it will not produce:

  • Reviews nobody wrote
  • Pricing it did not see
  • Features it cannot point at
  • A complaint with no page behind it, unless it is labelled unverified

If no source was provided, the fields stay thin and it says so. An empty analysis that admits it is empty is more useful than a full one you cannot check — and the temptation runs entirely the other way, because a rich-looking teardown is what people expect to receive.

Doing it by hand

Go to where the complaints are, not where the marketing is. Review sites, alternatives-to pages, support forums, the incumbent's own changelog comments. What you want is somebody describing the problem in their own words, unprompted.

Sort complaints into "will not fix" and "has not fixed yet". An incumbent that is enterprise-first will not build the thing your slice needs — that is a durable gap. A missing feature on their public roadmap is not a gap, it is a countdown.

Check whether the complaint has a buyer attached. Plenty of loud complaints come from people who were never going to pay. The complaint that matters is from someone already paying somebody.

Write down what you could not find. If you searched for a week and found no specific complaint, that is your answer, and it is the one people delete from their notes.

The short version

  • "Is it taken" is unanswerable in any useful way. Count of competitors tells you nothing about room.
  • The real question: is there a specific, repeated complaint the incumbent will not fix.
  • Three possible decisions — take them on, narrow to a slice they serve badly, or walk away. Silence usually means walk away.
  • Every complaint needs a source or an explicit mark that it has none, or the analysis is fiction with a confident tone.