Audience measurement — balancing test

This document is the legitimate interests assessment (LIA) for one narrow processing activity: measuring audience on vodster.app with Cloudflare Web Analytics. We publish it because we rely on legitimate interests (Article 6(1)(f) GDPR) rather than your consent — and when that is the case you deserve to see the reasoning, not just its conclusion.

It does not cover the mobile app or the consent-gated analytics (Google Analytics 4, PostHog) — those are described in our privacy policy.

1. Is the interest legitimate?

We run a where-to-watch catalogue covering several thousand films and series. We need to know which pages people actually open and where they arrive from so we can build on what helps and retire what does not. Without it we develop the site blind and waste work on pages nobody reads.

An operator measuring audience on its own site is a recognised legitimate interest — Recital 47 GDPR states plainly that a controller's interest can ground processing where it matches the reasonable expectations of the data subject.

2. Is the processing necessary?

We considered three alternatives and rejected each for a concrete reason:

• Consent-gated analytics (we already run it — GA4 and PostHog). It only measures people who click the banner. Search traffic usually leaves before it even decides about the banner, so this method systematically distorts the picture: it mostly shows already-engaged users and barely sees new ones. That is not a smaller sample of the same population — it is a different population.
• Server logs. They cannot tell a page view from a poster or stylesheet fetch, and they contain full IP addresses — processing more personal data, not less.
• Not measuring at all. That means running the site without knowing what works in it. We judged this disproportionate against the negligible intrusion described below.

Cloudflare Web Analytics is the least intrusive tool that answers the question: it stores nothing on your device and creates no identifier.

3. The balancing test

What we collect

• the page URL and the address you came from (referrer)
• country — derived from your IP address, which is never stored
• page load timings (Core Web Vitals)
• device type, browser, screen resolution

What we never collect or do

• No cookies and no localStorage — nothing is left on your device
• No identifier — your separate visits are never joined together
• No fingerprinting and no cross-site tracking
• No profiling, behavioural advertising or automated decision-making
• No sale or sharing of this data with anyone

Outcome

We see aggregate statistics, not people. We could not work out who you are from this data, nor tie it to your VODster account, even if we wanted to. The intrusion into your privacy is close to zero, and the risk of secondary use does not exist because there is nothing to re-use.

On the other side stands a real need to run the service on facts. We conclude that the legitimate interest prevails, and that the processing sits within the reasonable expectations of a visitor — a site operator counting visits is common and expected practice.

4. Retention and transfers

The data is held by Cloudflare, Inc. (USA) as a processor, in aggregate form, for about 6 months. That is materially shorter than the 13 months European data protection authorities cite as the ceiling for audience measurement. The transfer outside the EEA relies on standard contractual clauses and the EU–US Data Privacy Framework — details in our privacy policy.

5. Your rights

You have the right to object to this processing (Article 21 GDPR). Write to support@vodster.app — you only need to raise the objection, you do not need to justify it.

You can also block the measurement yourself, immediately: any script-blocking extension stops it without our involvement, and we use no techniques to work around such blockers.

6. Review

We review this assessment whenever the tool or the scope of collected data changes, and in any case at least once a year. Should the balance ever tip the other way, we will move the measurement behind consent and say so here.