Aglet

Investigate zero-result search recovery with state comparisons

An investigation should explain whether no results comes from retrieval, vocabulary, scope, inventory, or filtering. Replay the exact query, remove one condition at a time, and compare alternatives. Keep the empty message and available action in the evidence because recovery is part of the behavior.

Build a useful investigation brief

  1. Build the zero matrix

    List query, category, filters, result total, near matches, facet counts, empty copy, and recovery actions. Include hard zero, filtered zero, valid narrow zero, and an ordinary result. Preserve selection order for dynamic filters.

  2. Remove one boundary

    Change one factor—wording, category, filter, availability, or scope—and compare the result. Record whether a broader result helps the original task or merely hides the gap. Note when a suggestion changes the submitted phrase or only offers a link.

  3. Write the recovery brief

    State first divergent condition, user consequence, and competing explanations. Recommend one bounded empty-state or retrieval experiment with a pass condition. If the catalog cannot prove a near match, say so rather than inventing a recommendation.

What to carry forward

The investigation is ready when hard and filtered zeros have reproducible boundaries and a recovery action can be evaluated. Stop with one experiment or a clear unavailable state. Do not broaden results silently; preserve the query and disclose what changed.

Technical background: W3C guidance.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow