Autocomplete
Part of the Query Understanding series by Daniel Tunkelang.
Overview
Autocomplete predicts the rest of a query as a user types and presents completion suggestions in real time. It serves two purposes simultaneously: reducing the effort required to formulate a query, and steering users toward queries that are likely to return good results. Suggestions can be completions of the partial query, alternative related queries, or direct links to specific products or categories. The quality of suggestions depends on both relevance — do they reflect what this user probably wants? — and feasibility — will the suggested query actually return results? Suggestions are typically ranked by a combination of historical query frequency, correlation with successful search outcomes, and contextual signals specific to the current user or session.
Note: Article content behind Medium paywall — accessible at source URL with Medium account.