Docs
Console: Banner & A/B testing
Configure the consent banner and run statistically sound A/B tests to find the wording that converts best.
Banner configuration
Under Consent → Banner you configure what the visitor sees. Changes take effect on the next page load — no deploy needed.
| Setting | Options |
|---|---|
| Layout | bar (bottom strip) · box (corner) · modal (centre, blocks interaction) · wall (full-screen, forces a choice) |
| Theme | auto (follows prefers-color-scheme) · light · dark |
| Position | bottom-left · bottom-right · top (bar only) |
| Button labels | Accept all · Reject all · Manage preferences — each editable per locale |
| Show decline | Toggle whether the "Reject all" button is shown (off hides it; "Manage" remains) |
The CMP bundle reads the active config from
GET /v1/banner on every page load. You can also update it programmatically via PUT /admin/sites/:siteId/banner.A/B testing
Run a controlled experiment on banner copy, layout or button labels. Go to Consent → Banner → A/B tests → New test.
- Create variants — each variant is a full banner config override (different headline, layout, or button labels). Add 2–4 variants; traffic splits equally.
- Start the test — the CMP bundle rotates variants by visitor (deterministic, stored in a first-party cookie) so each visitor always sees the same variant.
- Read results — the console shows opt-in rate, 95% confidence interval, lift over control, and p-value per variant. Results update in real time.
- Pick a winner — once a variant reaches statistical significance (p < 0.05), promote it as the active banner config. Stop the test to freeze the winner.
Run tests in the TEST environment first to confirm the setup, then duplicate the test in LIVE. Opt-in rates differ between environments — don't compare across them.