Build a useful investigation brief
Build the counting matrix
For one query, list all products, selected filters, total hits, visible page, bucket values, and bucket counts. Repeat after each selection. Include a parent product with multiple variants and a single-record control so grouping cannot remain implicit.
Trace the first divergence
Compare filter application order, facet aggregation, pagination, cache state, and index timestamp. Check whether a count includes selected values, unavailable products, or variants the rows collapse. Record a deliberate distinction as evidence, not as an error.
Write the count brief
State the counting universe, failing boundary, affected filters, and evidence quality. Recommend a scoped fix or copy change with a pass condition. If freshness is the only gap, name the refresh observation needed before changing arithmetic.
What to carry forward
The investigation is complete when a representative count can be explained from query inputs through bucket output, or the first unknown boundary is named. Stop with a reproducible matrix. Do not promise exact counts when the system intentionally aggregates variants or uses a delayed index.
Technical background: Adobe developer documentation.
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