Resources · 27
Measure audiences while respecting privacy choices
Audit trackers, purposes, configuration and refusal journeys before publishing metrics.
· 20 min
What this guide helps achieve
- Map collection
- Assess conditions
- Test refusal
- Review access
Quick check
- What runs before a choice?
- Which purposes and recipients are configured?
- Does a claimed exemption meet all conditions?
- Do refusal and withdrawal stop relevant collection?
- How long is data accessible?
Step-by-step method
- 01
Inventory real behavior
Observe requests, browser storage and server flows before choice and after acceptance, refusal and withdrawal. Link every tool to a purpose and owner.
Deliverable: dated tracker and flow map.
- 02
Assess applicable rules
Compare actual settings with current rules, including the strict CNIL conditions for some consent-exempt audience measurement. A vendor claim does not establish an exemption by itself.
Deliverable: purpose and condition assessment per configuration.
- 03
Minimize collection
Remove unnecessary fields and review recipients, transfers, access and retention. Distinguish aggregate measurement from individual tracking.
Deliverable: minimal data schema and access register.
- 04
Test choices
Check that the banner and settings actually change requests and identifiers. Test a new browser, return visit, refusal and withdrawal.
Deliverable: QA evidence for each state.
- 05
Explain the method
State what metrics cover and miss, including refusals, blockers and differences among sources.
Deliverable: measurement note and comparison limits.
- 06
Review changes
Reassess after changing supplier, destination, tag, retention or banner. Obtain context-specific legal validation of the chosen basis.
Deliverable: review owner and schedule.
Management indicators
| Indicator | What it measures | First action |
|---|---|---|
| Inventory | Trackers with a purpose and owner | Investigate unknown flows |
| Choice | Refusal and withdrawal applied correctly | Fix configuration |
| Minimization | Justified parameters and access | Remove unnecessary fields |
| Review | Settings reassessed after change | Start a new verification |
Common pitfalls
- Calling a whole tool exempt without checking settings
- Inspecting the banner visually without network testing
- Mixing aggregate data with individual profiles
- Comparing audiences measured under different scopes
Frequently asked questions
Is a cookieless product always outside the rules?
No. The techniques and operations involving access or storage matter; the absence of the word cookie is insufficient.
Is a CNIL exemption automatic?
No. It depends on cumulative conditions and the actual service configuration.
How should a post-consent traffic drop be read?
Compare consistent scopes and explain unmeasured visits before attributing the change to visitor behavior.
Official references
Continue exploring
Related analysis
