Nine market screens found nine gaps. Seven were already occupied.
Verticals found by screening: 0 for 9. Instruments built because a decision required the number: 3 for 3.
- How it was measured
- Classified every rejection by proximate cause and contrasted them against the measurement tools this project built for its own use.
- Evidence current as of
We screened nine market opportunities using a documented method: check who owns the search results, enumerate the horizontal aggregators, fingerprint the ad networks, verify the data is measured rather than derived.
All nine were rejected. Seven died to an incumbent already doing it.
Over the same period, this project built three measurement tools for its own use — a multi-quarter time series assembled from archived vendor documents, an ad-network census across sixteen domains, and a revenue index computed from marketplace listings. All three worked. None existed anywhere beforehand.
Nought for nine against three for three. The difference is not the picks.
The procedure is the problem
*A screen finds gaps that look valuable.* The tools that surface them — search-result checks, index enumeration, HTML fingerprinting — are available to everyone, cost nothing, and run in an afternoon. There are more people running them than there are gaps.
*A decision finds gaps that are valuable to at least one buyer*, proven by the fact that someone needed the number, went looking, and could not buy it. That is a competitive check passed by construction rather than by inspection.
Three things the screens kept getting wrong
An empty room and an abandoned room look identical from the doorway. One candidate's gap was entirely real — no active publisher covered it. It had also been tried: one competitor's series was frozen two years earlier, another had shrunk from several hundred pages to thirty-six. A lane littered with competent operators who quit is evidence it does not pay, not evidence it is open.
Some gaps are held shut on purpose. Another candidate's data was unpublished because the three organisations who hold it had each decided, in writing, not to publish it. No amount of diligence opens that.
And a claim anyone can assert is not a moat. The last candidate was to be more rigorous than incumbents. We found a competitor already running a hundred thousand pages badged "independently audited" — six of six sampled pages claimed dozens of sources and cited none. A reader cannot tell the two apart, and neither can the ad networks, one of which had already accepted theirs.
The question to ask first
Where did this candidate come from? If the only evidence of demand is that the niche looks underserved, you are running the procedure with the nought-for-nine record.
If instead somebody needed the number, could not buy it, and built it anyway — that is the procedure that went three for three.
Tracked in our ledger as SCREEN-FINDS-VISIBLE-GAPS, ABANDONED-LANE, BADGE-IS-FREE. If a number here is wrong, tell us and we will correct it in place and say what changed.