Why it matters
Search introduces a second view of product data. It needs a tenant boundary, a freshness contract and a way to handle records that change or disappear.
Search becomes a product feature when users expect to find their own work quickly. Relevance, permissions and index freshness matter as much as the search box.
Decisions to make
- Identify the questions search must answer before choosing a search engine.
- Define permission filtering and result visibility.
- Choose how quickly updates must appear.
- Choose which records are searchable and which fields affect ranking.
- Define how fast edits and deletions must appear in results.
Build checklist
- Index only fields needed for search.
- Reconcile the index after missed events.
- Test permission changes, deleted records and large tenants.
- Apply authorization before returning results, including counts and snippets.
- Test reindexing, stale records and typo tolerance with real queries.
Common mistakes
- Treating a search index as an authoritative copy of customer data.
- Indexing private data without an access filter.
- Treating a development dataset as evidence of useful ranking.
Tools & references
4 starting points to compare or study. Check fit, maintenance, terms and current capabilities before choosing one.
