Is the denied cohort's conversion rate real?

Not necessarily. Denied visitors are often under-measured, so their rate may reflect missing data rather than behaviour.

Where do I get an export with consent status?

From a tool that records consent with each event, such as a server-side setup, a CMP export joined to events, or your own logs.

No. The file is parsed for the result and findings expire after 7 days.

What does a large gap mean?

That consent strongly shapes what your reports show, so channel and funnel comparisons should account for it.

What should I check next?

Whether consent is set before tags fire, with the Consent Race Condition Checker.

Free tool by Clarvivo · no login

Measure the analytics impact of consent status

Upload an export with consent and event columns to compare visible event rates without mistaking denied analytics for missing business activity.

Consent and event export

Upload a CSV with consent status and event name columns. We compare the cohorts without treating denied-consent analytics as business truth.

This tool answers: How much does consent status change the events visible in my analytics export?

Enable JavaScript to run the check — the diagnosis itself is free and needs no account.

Why this happens

Consent banners change what your analytics can see. Visitors who decline are often tracked in reduced form or not at all, so conversion rates, channel performance and funnel metrics are calculated from the people who said yes.

That matters because consenting and declining visitors aren't the same. If consent rates differ by channel, device or country, reports can show differences that are really just consent differences. Decisions made from those reports optimise for the consenting audience.

This analyzer compares conversion rates between consent cohorts in an export you upload, so you can see how big the gap is and whether consent should be part of how you read your numbers.

Next checks

These answer the questions this one tends to raise.

How this check works

So you can trust the result — and your developer can verify it.

Upload a CSV with consent status and events
The file needs a consent status column (granted or denied, or similar) and an event or conversion column.
We split the export into consent cohorts
Rows are grouped by consent state. If only one cohort is present, we say so, because there's nothing to compare.
We compare conversion rates
Conversions per row are calculated for granted and denied cohorts, and a material difference is reported with both rates.

Common causes and fixes

Denied visitors aren't measured fully: Without consent, tags send reduced or no data. Treat denied-cohort analytics as incomplete, not as business truth.
Consent rates differ by channel: Channels with lower consent rates look worse in analytics. Compare channels against store or CRM data before cutting them.
Banner design affects behaviour: Banners that block the page can increase bounce among people who decline.
Modelled data not enabled: Consent Mode with modelling can fill some gaps in Google's reports, if your traffic meets its thresholds.

Frequently asked questions

Is the denied cohort's conversion rate real?: Not necessarily. Denied visitors are often under-measured, so their rate may reflect missing data rather than behaviour.
Where do I get an export with consent status?: From a tool that records consent with each event, such as a server-side setup, a CMP export joined to events, or your own logs.
Is my data stored?: No. The file is parsed for the result and findings expire after 7 days.
What does a large gap mean?: That consent strongly shapes what your reports show, so channel and funnel comparisons should account for it.
What should I check next?: Whether consent is set before tags fire, with the Consent Race Condition Checker.

Do not let partial analytics become the whole story.

Clarvivo connects browser evidence with downstream source-of-truth data.

Connect measurement to revenue

What this tool reads, and what it keeps

Your export is parsed to produce the diagnostic and is not stored as a file.
Only aggregate columns and redacted row references are used in the result.
Findings expire after 7 days and are never used to train models.