Aglet

Investigate search result snippets with query-to-context traces

An investigation should find where useful context disappears between a query match and the rendered result. Replay representative content types, compare matched fields and excerpt windows, and inspect narrow rendering. Keep a destination check so a snippet can be judged against what it promises.

Build a useful investigation brief

  1. Build the snippet matrix

    List query, result type, title, matched field, excerpt source, highlight, truncation, viewport, and destination heading. Include one content type with a good excerpt and one with a buried match. Record whether the query term is visible or merely implied.

  2. Trace the first loss

    Compare source field, excerpt selection, normalization, rendering, and responsive truncation. Vary one content type or viewport at a time. Record when a snippet remains lexically correct but loses the context needed to choose, separating rendering from retrieval.

  3. Write the context brief

    State essential context, first loss boundary, affected templates, and competing explanations. Recommend one bounded field or layout experiment with pass criteria. Do not claim the result is irrelevant when the evidence only shows an incomplete excerpt.

What to carry forward

The investigation is complete when the first context loss is reproducible from match to destination and the affected result scope is clear. Stop with one bounded experiment. Keep content-type and viewport limits visible so a fix does not overpromise across all snippets.

Technical background: Digital.gov 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