# Consent-aware website analytics: what to check before trusting the numbers

- Link: https://codexcollective.co.uk/blog/the-dawn-of-the-post-cookie-age-a-10-billion-challenge-for-the-advertising-industry/
- Published: 2023-08-19T19:25:00+01:00
- Author: Codex Collective

Website analytics can only be interpreted properly when you understand how consent
choices affect collection. Replacing third-party-cookie predictions with another
technology does not remove the need to define what is collected, test the implementation
and explain its limits.

## Separate the banner from the implementation

A consent interface records choices. Tags, plugins and embedded services then need
to behave according to the organisation’s agreed requirements. A banner appearing
on screen does not prove that every script respects the selection.

Inventory analytics and marketing tags, including code added outside Tag Manager.
Record the owner and purpose of each integration.

## Test the states visitors actually encounter

 * A new visit before a choice has been made.
 * Analytics refused and a subsequent page navigation.
 * Analytics accepted and the important enquiry journey completed.
 * A changed preference and a returning visit.
 * Embedded forms and third-party journeys.

Inspect observed requests and storage rather than relying only on the visual state
of a toggle. Record the conditions and the expected result before testing.

## Understand what consent mode does

Google’s [consent mode documentation](https://developers.google.com/tag-platform/security/concepts/consent-mode)
describes signals that adjust Google tag behaviour. Configuration choices affect
collection and modelling. The implementation should follow the business’s agreed
requirements and be checked in context.

Consent mode does not obtain a visitor’s choice by itself and is not a certificate
of legal compliance. Technical configuration and privacy decisions have different
owners, who need to coordinate.

## Make reports honest about coverage

Do not label every unobserved journey as lost demand or treat a change in collection
as a change in marketing performance. Document consent-platform releases and tag
changes alongside reporting dates.

If analytics and operational enquiry totals differ, investigate definitions and 
collection before deciding either source is wrong. A confirmed submission and a 
qualified lead are different outcomes.

## Keep the setup maintainable

Use an event dictionary, tag inventory and repeatable test record. Retest after 
changing forms, templates, embeds or the consent platform.

For implementation support, explore [Analytics & Tracking](https://codexcollective.co.uk/analytics-tracking/).
For wider website diagnosis, our [Core Web Vitals guide](https://codexcollective.co.uk/blog/core-web-vitals-mistakes/)
explains how scripts can affect visitor experience as well as measurement.