Wisdom of crowds
Also called the wisdom of the crowd.
The wisdom of crowds is the idea that combining independent judgments from a diverse group can produce a more accurate estimate or decision than relying on one person. It works best when individual errors offset rather than repeat the same bias.
How it works
Ask people for their judgments separately, then combine their answers using a rule chosen in advance: an average, median or vote.
The group needs relevant information, not just lots of members. If everyone copies the first answer or uses the same faulty source, adding people does not fix the error. Discussion before collecting answers can turn separate judgments into one shared guess.
Four forecasts beat any single forecast
In this hypothetical example, four employees independently forecast next month's orders. The actual result later turns out to be 100.
| Forecaster | Predicted orders |
|---|---|
| A | 70 |
| B | 90 |
| C | 110 |
| D | 130 |
| Group average | 100 |
The arithmetic is (70 + 90 + 110 + 130) ÷ 4 = 100 orders. Nobody gets the answer right individually, but their errors cancel out.
This illustrates the mechanism, not a guarantee. If everyone overestimates demand, the average will also be too high.
For a startup, repeated investor feedback deserves the same check. Several investors repeating an objection are not necessarily independent signals if they all heard it from the same lead investor. Collect reactions separately before treating agreement as evidence.
Questions people ask
What is the wisdom of crowds theory?
James Surowiecki popularized the idea in his 2004 book The Wisdom of Crowds. He describes four conditions for a wise crowd: diversity of opinion, independence, decentralized knowledge and a way to aggregate judgments. The claim is conditional, not that a majority is always right.
Is wisdom of crowds the same as groupthink?
No. Wisdom of crowds depends on preserving different judgments before combining them. Groupthink suppresses disagreement to maintain consensus. A unanimous meeting can therefore be weaker evidence than a spread of independently collected estimates.
