Sample

Ask which slice of the available traces actually got captured, and which got left out: whose cases, which events, what time window, and what quietly went missing. A true signal read off a lopsided slice still points the wrong way.

Ask which slice of the available traces actually got captured, and which got left out: whose cases, which events, what time window, and what quietly went missing. A true signal read off a lopsided slice still points the wrong way.

Example

In the satisfaction drop, you check the slice: the email survey went to all 4,000 customers, while the in-app popup only reached the 900 who logged in that week. Quiet, mostly-happy occasional users never appeared in the 71%.