Note 01
A QA capture is not traction.
A screenshot can look like a product. A product can look like a business. The gap between those three things is where most company sites quietly lie.
There are three layers between a build and a business, and they are easy to blur on purpose.
Surface. A screen exists. It renders, it navigates, it survives a narrow viewport. A capture proves this layer and nothing above it.
Flow. A person can complete a sequence: discover, request, agree. This is harder than a surface and much easier than demand. It proves the product is coherent, not that anyone wants it.
Demand. Strangers come back without being asked, and some of them pay. Only this layer is traction.
What the ledger is allowed to say
Our evidence page records surface and flow claims with the build version attached. It also records, in the same row, what has not happened yet: public availability, real users, repeat usage, transactions. The second column is not modesty. It is the part that makes the first column readable.
If a row says “portfolio and discovery surfaces render and navigate”, a stranger can ask to see it, and we can open the same build. If a row said “users love it”, there would be nothing to open.
The test I use
Before a claim goes on the site, I ask one question: can this be reopened? A capture, a build number, a gate sheet, a signed agreement — yes. A feeling, a compliment from a friend, a chart with no axis labels — no.
The test is uncomfortable because it removes the most flattering sentences. It is also cheap: it takes ten minutes to apply, and it prevents the worst conversation in any diligence process, which is “show me”.
What would change the label
Labels on this site are not permanent. The moment a venture has a cohort that returns, or a paid transaction that covers its cost, the row changes and the build log gets an entry. Until then, the label stays where the evidence is.
A capture is a promise that the product exists. It has never been a promise that the market cares.