Why the first result is clicked regardless of relevance

Researchers swap the top two results in a search engine without telling users. Click-through follows the position, not the content: whatever now sits first receives most of the clicks. Users report choosing the result that looked most relevant, and the result they chose is the one that moved.

What everyone sees

Position bias is often described as laziness, which is unfair and unhelpful. Assessing relevance requires reading several snippets and modelling which page answers the question — genuine cognitive work. Following rank requires none. The searcher is not avoiding effort out of indifference; they are using a signal that is usually correlated with what they want.

What is actually happening

Joachims and colleagues used eye-tracking to show that users examine the top results far more and that swapping ranks moves clicks with position. Pan and colleagues found that participants trusted rank over relevance even when the top result was manipulated to be a poor match. Gigerenzer and Goldstein’s work on fast-and-frugal heuristics explains why: a single cue that correlates with the criterion can outperform integration of many cues under time pressure, so rank-following is a rational default that becomes exploitable the moment ranking is controlled by an interested party.

Why it stays hidden

The substitution hides because the searcher generates a justification after the click. Having chosen the first result, they can describe why it looked apt, and the description is sincere. What they cannot observe is the counterfactual: the same account would have been produced for whichever result occupied the position, which is exactly what the swap experiments demonstrate.