See what your analytics payload actually carries.
Parse the events. Trace identifiers across UTC days. Keep the product signal without waving away the privacy evidence.
WAITINGNO ROWSLoad or edit the bundled CSV--
Parse the events. Trace identifiers across UTC days. Keep the product signal without waving away the privacy evidence.
Run the audit to produce a finding.
A clean result means only that this CSV contains no populated cookie/IP/user-agent/email cells and its anonymous IDs rotate under the stated grouping rule.
The bundled local CSV has eight events, four page views and two purchases. Two divided by eight is twenty-five percent of events, not conversion of unique people. The two purchase revenues, forty-nine and ninety-nine, sum to one hundred forty-eight. Common one-pixel-per-dollar bars measure forty-nine, ninety-nine and one hundred forty-eight. The browser parses supplied values; it does not inspect actual network payloads or prove any sales happened. In the bundled clean sample, both accounts have disjoint anonymous identifier sets on adjacent UTC days, so two out of two pass the source rotation rule, one hundred percent. The genuine counterexample gives account A a persistent identifier and adds a cookie and I P cell. One of two accounts passes, fifty percent, and exposed cells rise from zero to two. Rotation bars use four display pixels per percent, widths four hundred and two hundred. The eight events and one hundred forty-eight revenue stay unchanged. This rule is a technical CSV test, not consent or legal compliance. Consider one account across three UTC days with identifiers A, B, then A again. The source checks two adjacent transitions; both have zero intersection, so it counts that account as rotating, one out of one, one hundred percent. There are only two distinct identifiers across three days: bars at one hundred pixels per count have widths two hundred and three hundred. This shows why a passing result does not prove identifiers are never reused. A single-day account cannot pass. Sensitive cells are trimmed for exposure counts, while a spaces-only field can still label its row exposed. Editing or a failed audit keeps the previous downloadable report; reloading loads the default CSV instead of restoring the saved report.