Any refinement a user selects — grouping, sorting, or filtering — triggers a new search request, unless that exact search has already been cached client-side.
This is why request volume usually exceeds what you'd predict from visitor numbers alone. A single user session on a faceted results page might look like this:
- Page load, empty query → 1 request
- Select a category facet → 1 request
- Add a price filter → 1 request
- Change sort order → 1 request
- Page to the next set of results → 1 request
That's five billable search requests without the user typing anything at all.
Facets can multiply operations within a request
If you use disjunctive faceting, a single user query needs more than one query internally, the engine has to compute counts for the unselected values in each disjunctive facet group. The number of operations per request therefore rises with the number of disjunctive facets you use.
On plans billed by request this doesn't change your cost. On legacy plans billed by operation it does. See How Algolia counts search requests, operations, and records.
How refinements appear in Analytics
Your Analytics dashboard will not show five searches for the session above. Requests that share a user ID and fall within a 30-second window are aggregated into a single reported search, so a sequence of refinements typically collapses to one. See Why are the search request numbers in my Billing and Analytics pages different?