Personalization logs

Every time a personalization point or recommender runs, Personalization keeps a record of the decision. Use that record to confirm what actually served, time each stage, and debug a result that looks wrong — without re-running the flow live.

What a decision record tells you

What's captured per decision run
Category	What it tells you
Decision identity	Which point, decision and campaign fired
Content served	The recommender output or content item that was returned
Stage timing	How long each step of the decision took
Outcome	Whether the decision completed, used a fallback, or hit an error

Debug and audit with the log

Using decision records

Filter records to the specific point or decision you're investigating
Compare stage timings to find which step is slow
Check the outcome before assuming the recommender is wrong — a fallback or error explains an unexpected result
If expected records are missing, confirm the upstream connector/data stream feeding Personalization is still active

For the API call that produces a decision, see Call the Decisioning API. For an aggregated, dashboard-level view of the same pipeline, see the Pipeline Intelligence dashboard.

Discussion

Comments are not configured for local development. They appear once the site owner connects GitHub Discussions (Giscus).