Most organisations decide by getting comfortable. You bring an idea, people react, and it moves forward once enough of the room is nodding. Objections get sanded off. The version that survives is the version nobody minds.
Sequoia Capital ran an experiment on whether that works. On their own capital. For thirteen years.
A 13-Year Experiment on Their Own Money
Since 2014 they have tracked deal outcomes against how the partnership actually voted. Partners score conviction on a 0–10 scale rather than voting yes or no, and the co-stewards who run the firm get no tiebreaker weight.
The question they were testing: do consensus-backed investments outperform conviction-backed ones?
The Finding
It does not matter whether the vote was unanimous or split.
The only variable that predicts outcome is whether at least one partner has genuine conviction — an 8, 9 or 10. Five people at 7 is not a good sign. It is the absence of the signal.
The Vote That Was Nearly Unanimous — Against
In 2019 a brand-new partner pitched SpaceX at a $20B valuation. This was before Starlink. The menu at the time was software: Zoom had just IPO'd, and Tesla and Nvidia combined were worth about what Salesforce was worth alone.
The partnership was close to unanimously against it.
It went ahead on one person's conviction, and became one of the best returns in the firm's history.
A committee cannot produce a 9. Averaging is the one operation guaranteed to destroy an outlier.
Why Averaging Destroys Outliers
This is arithmetic, not culture. If a decision needs broad comfort, then anything genuinely non-obvious fails by construction — because non-obvious is precisely the thing most of the room cannot see yet.
Consensus is a good filter for avoiding disasters. It is a terrible filter for finding outliers. Most organisations use it for both and then wonder why nothing remarkable comes out.
There is a second variable, and Sequoia is explicit that it is separate: courage. Conviction is believing. Courage is continuing to push when it costs you something personally. One without the other produces nothing.
What To Do On Monday
You do not need a venture fund to use this.
- Stop asking "does everyone agree?" Ask "is anyone here a 9 on this?"
- Score, do not vote. Yes/no hides intensity. A 0–10 scale surfaces it.
- Protect the person who is sure. The default social gradient in every meeting runs against them.
- Treat "nobody objects" as a warning. It usually means nobody cares.
The uncomfortable implication: if nothing in your pipeline has a single passionate advocate and several skeptics, you are not being disciplined. You are being average, efficiently.
Source: Alfred Lin & Pat Grady (Sequoia Capital) on Bloomberg Tech, 2026.
