An app appears in a trend chart, a founder’s post and a competitor newsletter. It feels as if three independent sources have noticed a breakout.
Then you discover that the newsletter cites the post, and the post cites the chart. You have one observation wearing three outfits.
Draw the evidence chain
For each material claim, identify the original source where possible. Record who measured it, who reported it and who repeated it. A secondary explanation can be useful without becoming independent confirmation.
This matters with estimated revenue, downloads and growth claims. Two products may use related inputs or one may licence data from the other. Agreement can reflect shared methodology rather than separate measurement.
Combine different questions
Stronger research often combines evidence types that address different parts of the opportunity. A dated market observation may suggest attention. Reviews may reveal a repeated problem. Interviews may clarify switching costs. A paid pilot may test willingness to pay.
None of these alone proves the whole business, but together they can support a sensible next step.
Keep disagreement visible
If one source suggests growth and another shows declining interest, do not automatically discard the inconvenient result. Compare time windows, countries, platforms and definitions. The apparent disagreement may be explainable, or it may reveal uncertainty you should retain.
For a hypothetical niche-selection exercise, write a short evidence map before scoring the opportunity. Put unknowns beside the claims they affect. If a central assumption rests entirely on one opaque estimate, make the next experiment inexpensive and reversible.
The objective is not to collect enough citations to make a confident paragraph look respectable. It is to understand which conclusions the evidence can carry.
Three different kinds of evidence can be valuable. Three copies of the same number are still one number.