Consent is a visitor’s choice, and it changes what a report contains. It does not change what happened.
Two sources, two behaviours
Search Console counts clicks from search results. It is Google’s own data, and consent does not touch it.
Analytics counts sessions and behaviour on your site. Consent decides how much of that you will see.
A worked month
Brightwater Dental, September 2026.
source measure value note
search impressions 1,860 consent does not apply
search clicks 41 consent does not apply
search click rate 2.20% 41 / 1860
analytics sessions 118 consent applies
analytics engaged sessions 61 52% of sessions
analytics generate_lead events 9 consent applies
form inbox stored enquiries 7 no tag involved
the consent gap
41 clicks from search, 118 sessions across all channels.
The gap is not a problem: they count different things.
what the gap hides
An unknown number of visitors declined analytics. Their reading,
bouncing and clicking is invisible. The form inbox is the one
number that survives a decline, because it is a stored record
on your own server.
The plan that survives the gap
- Report search clicks as measured, with the source named.
- Report sessions as consent-dependent, in the same table.
- Never derive a conversion rate from a denominator you cannot trust.
- Keep the form inbox count as a floor for real enquiries.
- Say which numbers carry a known undercount.
Acceptance: the setup is honest when
- the tag collects nothing before a decision is recorded
- the report names which numbers depend on consent
- the inbox count is shown beside the event count
- no percentage mixes a consent-free numerator with a consent-bound denominator
- the consent rate itself is visible, so a shift is explainable
Troubleshooting
| Symptom | Likely cause | What to do |
|---|---|---|
| Sessions fell, clicks did not | Consent rate dropped, or the banner changed | Compare the consent rate week to week |
| Analytics shows zero for a day | The tag was blocked or removed | Check one page’s HTML for the tag |
| Events exceed stored enquiries | Phone clicks, or a second form | Name the difference in the report |
| The client asks for the true number | Nobody has it | Report the floor, and say what is missing |
| Every percentage moved at once | A denominator changed | Fix the denominator before reading ratios |
A worked consent change
The banner changed on 12 October, from a single accept button to accept and decline. Two weeks either side, on the same site.
period A, 2026-09-28 to 2026-10-11, 14 days
sessions 64
engaged 33 51.6%
key events 5
period B, 2026-10-12 to 2026-10-25, 14 days
sessions 41
engaged 27 65.9%
key events 4
search clicks over the same two periods
A 22, B 21
reading
sessions fell 36%, and search clicks did not move.
the visitors did not go anywhere: the tag simply sees fewer of them.
the engaged share rose because the people who accept tend to read more.
That is the shape to expect whenever a banner changes. Report both periods with the banner change noted, and refuse to call it a traffic drop.
Limits
No tag recovers a session a visitor chose not to share.
Where a decision needs precision, use the source that does not depend on consent. Say which one you used. On a small site that is usually the search click count and the stored enquiry count.