Broad and Ambiguous Search Queries

Source: https://medium.com/@dtunkelang/broad-and-ambiguous-search-queries-1bbbe417dcc Author: Daniel Tunkelang

Summary

Exploration of the challenge that broad and ambiguous queries pose for search systems. These queries have multiple valid interpretations and require diversity/disambiguation strategies rather than precision optimization.

Types of Broad/Ambiguous Queries

  • Broad: “shoes” — clear topic but enormous scope
  • Ambiguous: “apple” — genuinely unclear intent (fruit vs. company vs. color)
  • Vague: “good stuff” — undefined quality criterion

Why Standard Ranking Fails

Relevance algorithms optimize for the “best” single interpretation. For broad queries:

  • The “best” interpretation may serve only a fraction of users
  • Result homogeneity leaves many users unserved
  • Users who don’t see their intent represented will reformulate or leave

Strategies

  1. Diversify results — represent multiple interpretations in the result set
  2. Clarification queries — prompt user to narrow intent (risky: adds friction)
  3. Personalization — infer likely intent from user context
  4. Faceted Search — let users self-select their interpretation dimension

People