Platform risk is the third most common way our gate kills an idea, not the first

Platform risk gets endless warnings and almost no denominators. I have one for my own gate, and the number that surprised me was not the share — it was the rank.

Across 706 rejected candidates, platform dependency was named in 19.8 percent. That puts it third, behind the missing documented problem at 438 and hands-on support load at 187.

Third. The thing that gets the essays and the cautionary tales is not the thing killing most ideas. What kills most ideas is that nobody could find anyone complaining about the problem.

Then the part that made me change the rule. Of the 140 candidates that named platform risk, it was the only reason in 15. I went and read those 15 — Shopify apps, Discord monetisation, Upwork tooling, property-listing widgets, Instagram schedulers. Every one reaches a buyer through a marketplace.

Now put that next to a founder profile that says no calls, async only. For that person a marketplace is often the ONE channel that delivers a paying customer without a conversation. I was killing ideas for depending on the only distribution my own product's constraints left them.

So platform risk became the single soft check: it costs a candidate its rank, not its life. Revived ones carry the kill reason as a visible flag, sit at the bottom of the ranking, and fill leftover capacity only. Platform risk plus anything else stays dead — that is 125 of the 140.

Two things I would not have thought to say before a review caught them. Pinning a revived candidate's score to zero was not enough, because zero is a legal score for a candidate that really passed, and the tie-break was input order. And the prompt tells the model to score anything it kills at zero, so a revived candidate has no honest merit number at all — inventing one would have been the dishonest fix.

It also does not override the founder. If your profile says the dependency is a dealbreaker, the hard kill stands.

Measured on 23 real runs, all my own accounts, on 2026-09-06. It describes my gate, not a market: https://whittleos.com/guides/platform-risk-startup