Personalization
Part of the Query Understanding series by Daniel Tunkelang.
Overview
Personalization adapts search results to individual users based on their history, preferences, and behavior. The same query from different users may warrant different results — past purchases, browsed categories, preferred brands, and explicit profile settings all provide signal about what a given user is likely to want. The challenge is weighting these signals appropriately: personalization should inform but not override explicit user intent, and over-reliance on historical behavior can trap users in a narrow view of what results they’ll see. New users present a cold-start problem with no history to draw on, and shared devices complicate attribution of signals to individuals. Privacy regulation adds further constraints on what data can be collected and how long it can be retained.
Note: Article content behind Medium paywall — accessible at source URL with Medium account.