This article details the definition and calculation method for each indicator displayed on your Statistics page. It follows the page organization from top to bottom so you can easily find what you're looking at.
​

Keep in mind All indicators are calculated strictly over the analysis period you select. What happens outside this period is never included in the calculation.

Before reading your figures, you define at the top of the page what you want to analyze. This choice applies to all indicators and charts displayed below.

These three indicators measure how your visitors react to your banner. Breakdown by device (desktop or mobile) remains accessible via "See more".
​
​

​

The proportion of visitors able to interact who actually clicked on the banner.
​

An interaction refers to any action on the banner: Accept, Refuse, Partially accept, Sign in, or Subscribe. Only interactions performed on the same banner configuration are counted.
​

A visitor is considered "able to interact" in two cases: they are new (they have never interacted before), or the banner was presented to them at least once during their visit. Details of this logic are explained in the FAQ.
​

Interaction rate = (number of interactions ÷ number of visitors able to interact) × 100
​

The proportion of visitors who consented among those who interacted with the banner. Consent corresponds to acceptance, total or partial. A refusal does not count as consent.

Consent rate = (number of consents ÷ number of interactions) × 100
​

The proportion of visitors who saw the banner and ultimately consented. This is the most complete performance indicator, as it combines the propensity to interact and the propensity to accept.
​

Opt-in rate = (number of consents ÷ number of visitors able to interact) × 100
​

Do not confuse consent rate and opt-in rate. Both share the same numerator (consents) but not the same denominator. Consent rate is calculated among visitors who interacted; opt-in rate among all those who could see the banner. The opt-in rate is therefore mechanically lower.

These indicators describe what happens on your site, independent of consent.
​

​

The number of unique visitors to your site. Each visitor is identified by a unique token, stored in an axeptio_cookies cookie created on their very first visit.

A visitor who returns on the same device and same browser keeps their token. They remain a single unique visitor, whether they return once or ten times.

A new token is only generated if the cookie expires (default duration of 190 days, configurable per project), if the browser's local storage is cleared, or if browsing data is deleted.

The token is linked to a device + browser pair. The same person visiting from their computer and then their phone therefore counts as 2 visitors.
​

The number of visitors whose behavior suggests they are automated traffic. We always speak of "suspected" traffic, as no signal allows absolute certainty.
​

This traffic is excluded from all your rates (interaction, consent, opt-in, bounce), so they reflect the behavior of your real visitors. The identification method is detailed below.
​

The number of times your consent banner was presented to a visitor. Each opening is counted: if the banner reopens during the same session, each opening is added up.
​

A visitor "bounces" when they view only one page and leave in less than 5 seconds. Beyond 5 seconds, or if they view multiple pages, their visit is considered normal.

This rate helps you measure the share of very low-engagement traffic that didn't have time to truly explore your site.

Bounce rate = (number of visitors who bounced ÷ number of visitors) × 100

Below the indicators, several charts allow you to track your data over time, compare them, and read the trend. In opt-out mode, those covering consent adapt and only display acceptance and refusal curves.
​
​

​

The total volume of pages viewed on your site over the period. A URL change or page reload counts as a page view; simply switching the tab to the background does not count.

This chart helps you track your site's raw traffic and spot your traffic peaks and troughs.
​
​

​

The number of unique visitors over time, with a distinct curve for traffic suspected to be automated. You can distinguish at a glance your real audience from suspected bot traffic, and track your traffic trend.
​
​

​
​

The change over time in interactions with your banner (acceptances, refusals, partial acceptances). This chart lets you see if your consent performance is improving or declining, and measure the effect of a configuration or design change.
​
​

​

The distribution of bounces and their trend over time. You track, period after period, the share of your low-engagement traffic (single page viewed and departure in less than 5 seconds).
​
​

​

A table displaying, for each of your banners, the interaction, consent, and opt-in rates side by side. This is the most direct tool for comparing the performance of your different banners and identifying which one converts best.
​
​

​

A world map showing the geographic origin of your traffic. Beyond simple curiosity, this view is useful for understanding your international audience and anticipating which regulations (opt-in or opt-out) apply to your visitors.

This section only applies to projects with opt-out banners (regulations like CCPA in the United States or Law 25 in Canada). If your project doesn't contain any, you will only see the opt-in analysis described above.

Opt-out banners operate with inverted logic compared to opt-in. By default, data processing is permitted, and it's up to the visitor to act if they wish to object. Opt-in indicators (which measure active acceptance) are therefore not relevant here, which justifies a dedicated analysis area.

In this area, you will find:

Important: the opt-out calculation method differs from opt-in. The precise definitions of these indicators (opt-out rate, "Do Not Sell" rate, counting rules for automatic and manual opt-outs) are being stabilized and must be validated before publication. This section will be completed once these definitions are confirmed.

To ensure the reliability of your rates, traffic suspected to be automated is identified and then excluded from all rate calculations. It continues to appear in the "suspected bots" indicator and in the corresponding curve, to remain visible without skewing your performance.

A visitor is suspected to be a bot when identified as a robot based on their browser, when they generate an abnormally high volume of events, or when their behavior is characteristic of a bot (visiting a single page followed by immediate departure).

In case of doubt on a given day, the most "human" profile always wins: it only takes one moment during the day matching real visitor behavior for the entire day to be counted as such.

An interaction refers to any click on the banner (Accept, Refuse, Partially accept, Sign in, Subscribe). Consent refers only to acceptance, total or partial. A refusal is an interaction, but not consent.
​

Because their denominators differ. Consent rate is measured among visitors who interacted with the banner. Opt-in rate is measured among all those who could see the banner, a larger set. With the same numerator, the opt-in rate is therefore always lower.
​

In two situations: either they are new (or have never clicked the banner before), or the banner was presented to them at least once during their visit, even if they had already interacted in the past.

Because the banner can be redisplayed. The widget reappears notably when the visitor changes configuration, changes domain, when new regulation requires redisplay, or when their previous consent expires. As soon as the banner redisplays, the visitor rejoins the set of visitors able to interact.
​

The banner is not redisplayed to a visitor who has already consented in the past 6 months. Concretely, a visitor who interacted 2 months ago is considered to have already interacted, while a visitor who interacted 9 months ago is not: their consent is too old, and the banner will be presented to them again. Only interactions performed on the same configuration are considered.