How does Consent Mode affect ecommerce reporting?

Consent Mode passes a set of states to Google tags describing what the visitor agreed to. Separate signals cover analytics storage, advertising storage, whether user data may be sent for advertising purposes, and whether advertising may be personalised. A tag reads those states and adjusts its behaviour rather than being blocked outright.

That is the important distinction from simply not firing a tag. With consent denied, the tag can still send a request without identifiers, which tells Google that an event happened without saying who it belonged to. Those cookieless signals are what makes modelling possible.

The signals are also independent. A visitor can permit analytics and decline advertising personalisation, which produces reports where a session is visible in analytics and unavailable for remarketing. Understanding which signal is missing explains most confusing report behaviour.

In the European Economic Area, passing the advertising signals is required for personalised advertising features to work, and a store selling into those markets without them will see audience and measurement features degrade rather than fail loudly.

In a basic setup, tags do not load at all until consent is granted. Nothing is sent for visitors who decline, so those people are entirely absent from your data and there is no signal for modelling to work from.

In an advanced setup, tags load in a restricted state and send cookieless signals before consent is decided, then switch to full behaviour if consent is granted. That preserves a shape of the traffic even where identifiers are unavailable.

The reporting consequence is significant. An advanced implementation typically shows a smaller apparent gap because part of the missing behaviour is estimated, while a basic implementation shows a clean but incomplete picture. Neither is dishonest; they answer different questions.

Decide which you are running before interpreting anything, and record the date it changed. Most confusing year-on-year comparisons in a store's analytics turn out to be a consent implementation change rather than a change in the business.

Expect a step change on the day the banner or the implementation went live. Sessions, users and conversions can all move, and the size of the move depends on how many visitors decline, which varies enormously by market.

Expect a share of conversions to be modelled rather than observed. These are estimates of behaviour that occurred but could not be measured directly, and they behave sensibly in aggregate while never matching an individual order in your store's records.

Expect some breakdowns to be unavailable. Where the volume of data behind a segment is too small, reports can withhold detail to protect individual privacy, which shows up as missing rows rather than as an error message.

And expect smaller audiences. Lists built for remarketing only include people whose consent allows it, so a store in a market with a high decline rate will find its remarketing reach materially lower than its traffic suggests.

Set the default states before any tag loads, then update them when the visitor chooses. A banner that sets consent after the tags have already fired produces inconsistent data and, in some markets, a compliance problem as well.

Use a consent tool that genuinely integrates with your tag setup rather than one that hides a banner over a page while everything fires underneath. On Shopify, the platform's own privacy and consent features are designed to carry these signals through to the tags in the checkout and the storefront.

Then test it directly. Decline consent in a fresh browser and confirm that no identifiers are written, that events still arrive in the restricted form if you run an advanced setup, and that granting consent afterwards changes the behaviour as expected.

Finally, record the configuration and the date in the same place your team reads reports. Half of the confusion around consent data comes from analysts comparing periods across an implementation change nobody documented.