Cookies & Consent Rules
Transcript
Made automatically from the video's sound, so it may contain mistakes. Click a line to play from there.
Right, let's jump straight into episode three of this explainer. Today we're mapping out the real rules around cookies and consent, specifically tailored for product teams. If you're a product manager, founder, developer, or marketer, you are in the exact right place.
We're gonna translate these incredibly strict legal requirements into actionable architecture and UX design, cutting through the legal jargon so you can build with total confidence. Following up on episode 2's deep dive into PII and personal data, I'm going to assume you already have a solid grasp of the GDPR fundamentals.
And just as a quick reminder, this explainer provides general structural information, current as of November 2026, and is absolutely not legal advice. So please always consult your own counsel for specific issues. Here's our roadmap for today's session.
We'll quickly cover the scope beyond cookies, look at the strictly necessary exemption, dive into compliant UX design, and finish up with regulator nuances and QA testing. We're gonna start by tackling a massive product development myth right out of the gate.
I bet you've heard someone on your team ask, hey, if we aren't collecting personal data, we don't actually need a consent banner, right? It sounds totally logical, doesn't it? If we anonymise the data, surely the privacy laws just don't apply.
Well, let's take a look at what the rules actually dictate. Moving into our first section, let's see how the law legally views that device your user is holding in their hand. The crucial thing to understand here is that the e-Privacy Directive, specifically Directive 2002 58EC article 5.3, protects the user's terminal equipment.
This is mirrored in the UK by PECR Regulation 6, and it's a completely separate requirement from the GDPR. The law essentially treats a user's phone or laptop as part of their private sphere. So gaining access to it or writing new information to it is seen as an intrusion into their private life, completely regardless of what that data actually is.
And this brilliantly illustrates why calling it a cookie law is honestly a terrible misnomer. It demands prior opt-in consent for local storage, indexed DB, tracking pixels, device fingerprinting and mobile SDKs.
Under EDPB guidelines 22023 and the SNIL's mobile app recommendations, if you're writing to or reading from a device, the underlying technology just doesn't matter. The legal requirement for prior consent remains exactly the same. To put the final nail in the coffin of that myth we started with, the Court of Justice of the European Union made this definitively clear in the landmark Planet 49 case.
Protecting terminal equipment requires consent, even if literally zero personal data is collected. The mere act of accessing or storing non-essential information triggers the requirement.
So if the net is cast that wide, let's explore the exemptions that are actually available to us as builders. Under WP29 opinion 042012, you can safely bypass the banner for strictly necessary technical functions.
We're talking about session authentication, remembering a shopping cart or security tokens like CSRF protection. But here is the major catch. It must be strictly necessary to provide the service the user explicitly requested, not what's necessary for your business.
So non-exempt trackers like cross-site analytics or social remarketing tags, which actually create joint controllership under EDPB guidelines 8-2020 will always require prior opt-in. A really common question from developers is whether we can just bypass this whole e-privacy opt-in thing by claiming legitimate interest or contractual necessity under the GDPR.
It's a clever thought, isn't it? Trying to use one regulation to essentially sidestep another. Well, the courts have completely shut that down. The CJEU's Meta ruling, alongside the Digital Markets Act and recent EDPB guidelines, 1 2024, definitively prove that legitimate interests simply cannot substitute for Article 5 3 prior opt-in consent.
You can't just write a broad clause in your terms of service claiming tracking is a contractual necessity, and you definitely can't weigh your business interests above the user's privacy rights. Explicit consent is the only valid legal basis here.
Now what's really fascinating about this next section is how we have to translate these strict legal rules directly into our user interfaces. The European Data Protection Board took this so seriously that they actually formed a specific task force just to crack down on bad banner designs.
Their incredibly detailed 2023 report found that a massive chunk of the web was using deceptive design to force users into accepting cookies. They have now laid out exactly what regulators are hunting for. To achieve valid consent under GDPR Articles 411 and 7, your design must feature symmetric, equally prominent accept all and reject all buttons, right there on the very first layer.
If you hide the reject button behind a settings link or use pre-tick boxes, your consent is legally invalid. A user has to have a genuine free choice, and it must take exactly the same amount of effort to reject as it does to accept.
Furthermore, the FTC in the US, and specifically Articles 25 and 28 of the Digital Services Act, explicitly outlaw what we call dark patterns. You absolutely cannot assume scrolling down a page equals consent. You can't use deceptive button colours, like hiding an invisible grey reject button next to a glowing green accept button.
Overloading users with choices is now a direct, penalized breach of compliance. On top of that, regulators like the EDPB and the California Privacy Protection Agency mandate that withdrawing consent has to be just as easy as giving it. You cannot trap your users.
A stance that's heavily reinforced by recent EDPB opinions limiting those forced consent or pay models. Usually this means providing a persistent, easily accessible icon, like a little hovering shield in the corner of your site, allowing users to instantly revoke access without jumping through a dozen hoops.
Let's move to our final section and see how all of this feeds into your actual QA testing and national deployments. You really have to navigate national regulator nuances carefully here. For instance, France's CNIL, sheet number 16, offers a very narrow exemption, allowing specific, strictly configured first-party audience measurement tools to run without prior consent.
But across the channel, the UK ICO and most general EU regulators strictly require prior opt-in for all standard third-party analytics. You've got to know where your primary user base is and architect your systems accordingly.
For the product and dev teams watching, this is where the rubber meets the road. Your mandatory QA process is surprisingly straightforward. Open your browser DevTools, clear your cache, load your site and watch those network and application tabs.
You must verify that absolutely zero nonessential scripts execute and zero non-essential storage is set before an affirmative click on that accept button. If a single ad pixel fires while that banner is still on the screen, your implementation is broken.
Finally, you should absolutely be future proofing your platforms to detect browser level signals, like the W3C Global Privacy Control, specifically looking for the SEC GPC 1 header. Respecting these signals at the server level helps automate compliance entirely and provides a vastly superior, much less intrusive user experience.
To wrap up this explainer, I want you to consider how these incredibly strict EU and UK opt-in requirements we've just discussed will inevitably collide with the completely different opt-out landscape across the Atlantic, which is the exact provocative challenge we'll unpack in episode 4, the US picture.
Until then, keep building, keep testing those DevTools, and thanks so much for joining me.