Aglet

How to verify a slow search improvement without losing results

Search verification must prove that people reach the same useful results sooner. Reuse the original slow and fast terms, filters, account shape, and network profile. Measure the first useful result and final completion, then probe response ordering, pagination, and empty results for correctness.

Check whether the outcome improved

  1. Replay the original slow query shapes

    Use the exact terms, filters, sort order, dataset size, and interaction mode from the baseline. Record input-to-request, request-to-response, first result, final result, and total request count. Repeat enough runs to see variation and keep cache state explicit for each comparison.

  2. Check ordering and stale response handling

    Issue a slower broad term followed by a faster specific term, or type the original sequence quickly. Confirm that the final view belongs to the latest input and that loading feedback matches the active request. Compare result identifiers, order, counts, and pagination boundaries with the baseline.

  3. Exercise empty, broad, and filtered cases

    Verify no-match behavior, a broad result set, a narrow filter, and the next page or continuation. Confirm that the optimization does not omit records, duplicate rows, alter ranking, or hide an error behind an empty state. Test both explicit submit and incremental search when both exist.

What to carry forward

Classify search as improved, regressed, unchanged, or inconclusive with timing and correctness evidence attached. It passes only when representative slow terms reach useful results sooner, latest-input behavior remains correct, and empty or paginated states still communicate the right outcome.

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