Silent failures
A deployment breaks the banner script on Tuesday. You notice in next month's report. The Daily Interaction Rate chart would have shown you a cliff to zero the same afternoon.
Consent Analytics
Stop guessing what your cookie banner is doing.
Every banner impression, every click, every ignore — measured. CookieLet turns consent into a dataset you can read, broken down by country, device and banner layer, and backed by an auditable log.
No credit card
One script tag
GDPR & CCPA ready
Displays
397
Interactions
29
Ignores
368
Interaction rate
7%
Interaction vs Ignores
7 days1.7%engaged
- Interactions 1.7%
- Ignores 98.3%
Daily interaction rate
7 daysCountry comparison
Interaction rate Accept rate93%
of visitors never answer a typical banner — and every one of them is a measurement gap
3
segmentation lenses: country, device and banner layer
1
script tag to install — no tag manager gymnastics
100%
of decisions written to a timestamped, auditable consent log
The problem
A banner you can't measure is a liability you can't defend.
Most teams install a consent banner, watch their analytics traffic fall, and never find out why. Was the banner broken on mobile? Did it render in the wrong language? Did anyone actually click it? Without instrumentation, those are all the same invisible problem.
Regional blind spots
Your banner performs fine in one market and gets ignored in another. Aggregate numbers average that away. Country Analytics puts the two side by side.
Untestable design
You rewrite the banner copy and hope. Layer Analytics tells you whether the change moved decisions on the first layer or just pushed people into the preference centre.
Three screens
From a five-second glance to a forensic breakdown.
Consent Analytics is built as a funnel of detail. Start on the dashboard, drill into engagement, then segment until you find the cause.
My Dashboard
Health in five seconds
The landing screen answers one question first: is consent being collected at all? A live Banner Status chip sits above everything else, because if it isn't green, no other number on the page means anything.
- Banner Insights — views and interactions for the period, with the interaction rate calculated for you.
- Cookie Summary — total cookies and categories from the latest automated scan, with the scan date printed inline.
- Total Visit — a daily traffic line to sanity-check everything else against.
- Consent Trends — accepted, rejected and custom, as a share of decisions actually made.
Consent Trends
Last 7 days100%accepted
- Accepted 100.0%
- Rejected 0.0%
- Custom 0.0%
Read it in context. Consent Trends is a share of decisions, not of views. A perfect accept rate on a handful of clicks is a small sample, not a win.
Interactions
The engagement report
Four numbers that always reconcile, plus two charts that show you when things changed. This is the screen you open when analytics traffic drops and you need to know whether the banner is the cause.
Displays = Interactions + Ignores
397 = 29 + 368
// if these ever disagree, it's a data-integrity signal
- Interaction vs Ignores — the split, as a donut, with its own period filter.
- Daily Interaction Rate — day-by-day, so a broken deploy shows up as a cliff you can date.
Displays
397
Interactions
29
Ignores
368
Interaction rate
7%
Comparison
Who behaves differently?
The same events, sliced three ways. Each segment shows its display volume alongside two bars — interaction rate in amber, accept rate in blue — so you can see engagement and outcome in one row.
- Country — ISO-coded, volume-ranked. High displays with flat interaction usually means a language gap, not a design gap.
- Device — desktop vs mobile. A mobile rate well below desktop points at a viewport, fold or tap-target problem.
- Layer — the banner, the preference centre and the re-open control, measured separately.
Device Analytics
%10 mobile displays against 387 desktop is unusual for consumer traffic — the kind of anomaly that only surfaces once you segment.
Definitions
Every metric, defined once — so nobody argues about the denominator.
| Metric | Formula | Denominator | What it tells you |
|---|---|---|---|
| Display | count(render) | — | The banner was presented to a visitor. The base unit of everything. |
| Interaction | count(action) | — | Accept, reject, save preferences, or open a deeper layer. |
| Ignore | displays − interactions | — | Seen, not answered. Not a rejection — but not consent either. |
| Interaction rate | interactions ÷ displays | Displays | Whether your banner is being answered at all. |
| Accept rate | accepted ÷ decisions | Interactions | What people choose, once they choose something. |
| Accepted | "Accept all" | Decisions | Consent to every category. |
| Rejected | "Reject all" | Decisions | Consent withheld from all non-essential categories. |
| Custom | subset selected | Decisions | Granular consent set in the preference centre. |
Silence is not consent — and it isn't refusal either. Interaction rate is measured out of everyone who saw the banner. Accept rate is measured out of everyone who answered it. Two different denominators, deliberately. An ignored banner never inflates or deflates your accept rate, because it was never a decision.
Layer Analytics
The only view that tells you where in the banner people decide.
A consent experience isn't one screen — it's a stack. Measuring them together hides the story. Measuring them apart tells you exactly which surface to fix.
L0
The re-open control
The floating consent icon or "Cookie settings" link a visitor uses to revisit their choice. Under opt-in rules, withdrawing consent must be as easy as granting it — this layer is how you prove it's reachable. A permanent zero means nobody can find it.
0interactions
L1
The primary banner
The first thing a visitor sees, carrying the headline accept and reject actions. This is your volume layer — most decisions should happen here. Low L1 interaction relative to displays means the copy or button hierarchy isn't working.
19interactions
L2
The preference centre
The granular, per-category screen. Visitors who reach it are deliberate. High L2 volume with a lower accept rate isn't a failure — it's informed, granular consent, and it's the most defensible signal in the whole module.
10interactions
Watch the gradient. Accept rate should fall from L1 to L2 — that's normal. If it inverts, and people accept more in the preference centre than on the banner, your first layer is confusing enough that visitors are escaping into the settings screen to work out what's being asked.
In practice
What teams actually do with it.
Answer a regulator
Banner status confirms consent was being collected. Consent Trends shows the aggregate. The Consent Log produces the timestamped, per-visitor records behind it. Aggregate plus evidence, from the same source of truth.
Diagnose a traffic drop
Open Daily Interaction Rate. A vertical cliff to zero on a specific date almost always maps to a release. You get the date, then your release log gives you the commit.
Fix the mobile experience
Compare mobile against desktop interaction rate. A material gap sends you straight to banner layout on small viewports — buttons below the fold, a preference centre that won't scroll, tap targets under 44px.
Localise for a market
Find the countries with high displays and near-zero interaction, then check whether you're serving them a banner in a language they read. Translation is usually the cheapest interaction-rate win available.
Validate a cookie scan
Cookie Summary prints the scan date on the card. If it's stale, or the discovered cookies aren't all categorised, your banner isn't governing everything it should — and your accept rate is measuring an incomplete picture.
Prove withdrawal works
Non-zero Layer 0 activity is direct evidence that visitors can find and use the re-open control. It's a compliance metric, not a growth metric — its job is to exist, not to be large.
How it works
Installed in an afternoon. Reporting the same day.
Drop in the script
One tag, before your other scripts. CookieLet renders the banner, blocks non-essential cookies until a decision is made, and starts recording displays and interactions immediately.
Scan and categorise
The scanner discovers what's actually setting cookies on your site and sorts it into consent categories. Cookie Summary reports the count and the scan date so you know the picture is current.
Read, change, measure
Set your baseline. Change one thing — copy, placement, a translation. Watch the Daily Interaction Rate chart for a full week before you conclude anything. Repeat.
A word on optimisation
A defensible 60% beats an indefensible 95%.
It is trivially easy to raise an accept rate. Make the reject button grey and small. Pre-tick the categories. Bury the preference centre three clicks deep. Every one of those will move the number on this screen — and every one of them invalidates the consent behind it.
Consent Analytics is built to show you the real shape of visitor behaviour, including the parts that are inconvenient. Interaction rate and accept rate are kept separate for exactly this reason: so you can't quietly count silence as agreement.
Practices we'd rather you kept
- Give accept and reject equal visual weight on layer one.
- Leave non-essential categories unticked by default.
- Keep the preference centre one click away, not three.
- Make the re-open control visible on every page.
- Change one thing at a time, and give it a full week.
FAQ
Questions we get asked
Does an ignored banner count as a rejection?
No. Ignores are counted separately and excluded from accept rate. Functionally, though, under opt-in regimes an ignored banner means no consent — so non-essential cookies must not be set for that visitor.
Why do two charts on the same screen show different percentages?
Period filters are applied per chart, not per page. That's deliberate — it lets you hold a long-run baseline next to a recent window. Always check the filter chip before comparing two numbers.
Why is a country listed with zero displays?
The segment is known to the report — from an earlier period or from your configuration — but recorded no banner displays in the window you're currently viewing.
Does Consent Analytics store personal data?
The analytics screens present aggregate, anonymised metrics. Individual consent records, with the identifiers needed for audit, live separately in the Consent Log.
How often does the data refresh?
Metrics update on an ongoing basis as events arrive. Cookie Summary is the exception — it reflects the last completed scan, and prints that date on the card.
What counts as a good interaction rate?
It varies enormously by industry, region and banner design, so treat external benchmarks with caution. The number that matters is your own trend line: establish a baseline, make one change, and read the daily chart.
Can I export the data?
Consent records are available from the Consent Log for audit and regulatory response. For bulk export of aggregate analytics, talk to support.
Do I need to change my site to use it?
No. Consent Analytics is part of the banner. Once the CookieLet script is installed, displays and interactions are recorded automatically — there are no additional events to instrument.
Measure what your banner is really doing
Drop in the script, scan your cookies, and read your first opt-in numbers the same day.