Agree the scope before opening the browser

List the pages and journeys to test: the homepage, articles, contact forms, embedded media, campaign pages and any checkout or booking area. Include subdomains and externally hosted steps where relevant. Decide which browsers and mobile layouts represent a sensible sample for the site.

This is our suggested testing approach, not a statutory checklist. Its purpose is to avoid a common false assurance: testing one clean homepage while the advertising landing page uses a different template. Record the deployed version and test conditions so a developer can reproduce any finding.

Observe the first visit without touching the banner

Use a clean testing context with previous site preferences removed. Watch what loads before you choose anything. Inspect browser storage and network requests, noting the supplier, destination, timing and apparent purpose. A network call is evidence to investigate, not automatic proof of unlawful tracking.

The legal issue is whether non-exempt storage or access happens before valid consent. Some necessary operations or conditional exceptions may apply. Keep those classifications separate from the observations: first record what happened, then assess why it happened and whether the intended treatment is justified.

Reject, then continue using the site

The ICO expects refusal to be as easy as acceptance. Review the prominence, language and number of steps as well as the presence of a reject control. A visitor should not have to navigate a confusing sequence while acceptance takes one obvious click.

After rejecting, move between pages, refresh and use an embedded feature. Check whether disallowed activity begins on a later route. Some faults appear only when a component loads after navigation or when a tag manager receives a new event. Keep a short evidence trail showing the action and its result.

Test selective consent and later withdrawal

Where different non-exempt purposes are offered, test a selective choice. For example, enable one optional category and leave another off, then check the result. Labels should describe meaningful purposes rather than hide everything behind an undifferentiated preference.

Next find the route for changing the choice. Can a visitor reopen settings without deleting browser data or contacting support? Withdraw consent and test subsequent activity. Review what the mechanism promises about existing information as distinct from stopping future collection; those are different questions and should not be blurred.

Repeat on mobile and with accessibility in mind

A banner that works at a developer’s desktop size may cover essential content or push controls off a phone screen. Test text scaling, keyboard focus and the ability to reach both the choice and the explanatory information. An inaccessible control is not fixed by technically correct wording.

Check returning visits too. Is the saved choice respected? What happens when a preference expires or the purposes change? Do not invent a universal consent renewal interval: document the selected approach and why it fits the actual service. Look for unexpected resets that repeatedly pressure visitors to accept.

Report findings in a way that can be fixed

For each issue, describe the affected page, reproduction steps, observed behaviour, expected behaviour and supporting evidence. Distinguish a confirmed defect from an unanswered supplier question. Assign an owner and retest the exact failed journey after the change.

A useful final report does not simply say compliant or non-compliant based on a scanner score. Automated tools can help inventory technology but may miss conditional interactions or misclassify purpose. Combine them with manual journey testing and the supplier documentation. Prioritise any non-exempt activity before consent, activity after rejection and ineffective withdrawal, then address clarity and maintainability.

Make checking part of release discipline

Keep a small regression test for the key choices and run it after changes to templates, tag managers or integrations. Ask whoever introduces a new tool to identify its consent requirements before deployment. This turns the banner from a one-off purchase into a maintained control and makes later checks considerably less expensive.

Keep unresolved questions visible in the action log, with a named owner and the evidence needed to close them.

Sources and further reading

Primary guidance behind this briefing. Requirements depend on the facts; check the current guidance before making a decision.

This is general guidance, not a legal opinion on your circumstances or a guarantee of compliance. An individual assessment may be needed, particularly where sensitive information, disputed rights or regulatory action is involved.