Psychology and biases
Social proof
When unsure what to do, people copy what others like them appear to be doing, and treat the copying as evidence.
Robert Cialdini named it in Influence (1984); Solomon Asch's conformity experiments (1951 to 1956) are the founding evidence.
You are in a meeting where everyone nods at a plan you think is wrong, and you find yourself nodding. Later you half-believe it was fine. Or you pick the restaurant with the queue, the library with the most stars, the answer the other reviewers gave. Social proof is the rule 'what others do is probably right', applied without checking whether those others knew anything. Cialdini listed it as one of the main levers of persuasion. Asch showed its force with line judgements that had an obvious right answer: when the rest of the group gave the same wrong answer aloud, the real subject went along with them on 36.8 percent of trials, and three quarters of the 123 subjects yielded at least once.
It persists because it is usually a good heuristic. Other people often do have information you lack, and following them is cheap. The failure is a cascade: if each person copies the ones before, a group can converge on an answer nobody checked, and the apparent agreement of many is really the guess of one. Asch found the lever works in both directions. A single partner who gave the right answer cut the subject's errors to about a quarter of their usual rate, and even a dissenter who gave a different wrong answer freed subjects to make their own call. Product teams use the effect directly: 'most popular plan' labels, visible counts, and messages about what people like you did.
The core effect has replicated. Franzen and Mader re-ran Asch's line task in 2023 with 202 participants and found a 33 percent error rate, close to Asch's 36.8 percent; paying one Swiss franc per correct answer cut it to about 25 percent but did not remove it. The size varies, though. Bond and Smith's 1996 meta-analysis of strict replications reported an average error rate of 25 percent, and found higher conformity in collectivist cultures. The caution for a decision maker is to ask what the crowd actually knows. A hundred people copying one unverified claim is one data point, and a product that shows you counts is choosing which crowd you see.
Example
In a midsized, midpriced chain hotel in the American Southwest, Noah Goldstein, Robert Cialdini and Vladas Griskevicius randomly assigned 190 rooms to one of two towel-rack cards and recorded 1,058 occasions on which a guest could reuse a towel. The standard environmental appeal produced a 35.1 percent reuse rate; a card saying that most guests reuse their towels produced 44.1 percent. In a second experiment published with the first in 2008, a card saying that most guests who had stayed in this very room reused their towels reached 49.3 percent, against 42.8 percent for the other descriptive-norm messages combined.
Where it applies
Before adopting a tool because it is popular, find one person who chose it for a reason you can check. In a code review or a model evaluation, collect independent judgements before anyone sees the others', or the first opinion becomes everyone's.
Think about it
Which decision did you make this month mainly because people like you had made it, and what did they actually know?
Applied it to something real? Log it.
Test yourselfIn Asch's line experiments, what happened when one confederate gave the correct answer while the rest were wrong?
The subject's errors fell to about a quarter of the rate under a unanimous majority. One visible dissenter breaks most of the pressure to conform.
Related models
- Confirmation biasPeople look for, notice and remember evidence that fits what they already believe, and rarely test what would prove them wrong.
- Circle of competenceKnow where your real understanding ends, and keep your big bets inside that line.
- Network effectsA product becomes more valuable to each user as more people use it, which makes leaders grow faster and makes leaving hard.
- Feedback loopsA system's output feeds back into its input: balancing loops hold a level steady, reinforcing loops make change grow on itself.
Sources and further reading
- Solomon Asch, Opinions and Social Pressure, Scientific American 193(5) (1955)
- Noah Goldstein, Robert Cialdini and Vladas Griskevicius, A Room with a Viewpoint: Using Social Norms to Motivate Environmental Conservation in Hotels, Journal of Consumer Research 35(3) (2008)
- Axel Franzen and Sebastian Mader, The Power of Social Influence: A Replication and Extension of the Asch Experiment, PLOS ONE (2023)