To assess a Reject all button, read what the site promises and confirm what choice it stores. Then compare collection behavior with that promise over a defined window.
A site can make a narrower statement about refusing advertising cookies or explain that limited statistics continues. Preserve the exact explanation when reviewing the result.
Compare a promise with the observed result
Switch between a matched promise, a narrower promise, a mismatch, and unreadable consent. The continuing event needs its context.
- Banner promise
- Reject all blocks our optional analytics and advertising
- Saved choice
- Optional purposes denied
- Recorded window
- No example optional collection events; complete fixture capture
The fixture matches the specific promise for this window. It says nothing about later pages or server-side activity.
Next check: Repeat the same promise check on pages with different integrations.
Invented observations for teaching. Manual-debugger details illustrate evidence beyond CookieTrail’s current standard capture. Colors describe this example’s evidence or stated policy, not legal compliance.
Record the exact promise
Save the notice, category descriptions, and any explanation of what refusal affects. Distinguish declining advertising cookies, declining all optional purposes, and opting out of sale or sharing. Those statements don't necessarily describe the same operation.
The promise itself can matter. California's Healthline settlement resolved allegations including a consent banner that purported to disable tracking cookies but did not. This is a specific enforcement example, not a rate for websites generally.
- California Healthline settlement Allegations concerning opt-out enforcement and inaccurate banner behavior.
Verify the choice before measuring rejection
A disappearing banner is insufficient evidence of a stored refusal. Read a documented CMP API, stored-choice record, or other validated source. Record missing, conflicting, or unreadable states. Keep those cases out of an after-verified-rejection rate.
Use separate fresh sessions for rejection and acceptance. Accepting and then rejecting also exercises withdrawal, with old storage already present. Keep that transition separate from a first-time visitor refusing permission.
Compare collection with the stated scope
- Check cookie snapshots and, where available, the actual time of new writes. A cookie still present after withdrawal doesn't alone prove a new write.
- Distinguish downloading a script, attempting a request, and receiving a collection response.
- For Google requests, inspect consent signals and relevant storage behavior. Advanced mode can intentionally retain limited measurement.
- Compare windows of similar duration and retain attached-frame, endpoint, and other capture gaps.
Use the request-evidence diagnostic before diagnosing a continuing Google request.
Report promise outcomes separately
A useful report distinguishes a tested promise that matched, a demonstrated mismatch, a narrower disclosed policy needing review, and an inconclusive test. Don't pool these into a claim that a percentage of banners are illegal.
Count eligible independent sites with a verified refusal and the evidence needed for the named promise. Publish the number excluded for unreadable choices or incomplete capture. A general sale/sharing opt-out may require preference-signal and downstream evidence beyond browser events.
Review necessary-purpose claims for tools excluded from the optional controls.
CookieTrail can retain the standard regional paths and the saved-choice evidence available for that implementation. The resulting record gives your developer a concrete behavior to investigate and a matching baseline for a retest.