- A cookie banner is a tool for collecting consent. Whether you need consent is a separate legal question, and the answer depends on what your site stores or reads on the visitor’s device, why, and which national authority’s guidance applies.
- The ePrivacy rule (Article 5(3)) exempts storage that is strictly necessary for a service the visitor asked for. The Article 29 Working Party said in 2012 that first-party analytics cookies do not fall under that exemption, even though they carry limited risk.
- Authorities differ. France’s CNIL publishes conditions under which audience-measurement trackers can be exempt. Denmark and Ireland start from consent. Germany’s DSK declines to give a blanket answer.
- “No cookies” does not automatically mean “no consent”. The EDPB treats pixels and scripts that send device data back as within scope too.
- Do not trust a vendor’s “no banner needed” label. Audit what your site actually sets, then test it against your authority’s written conditions.
- This is general information, not legal advice.
“Do I need a cookie banner for analytics?” gets two kinds of answers online: a flat “no, our tool is exempt” from vendors, and a flat “yes, always” from consent-plugin sellers. Both skip the part that matters. The answer depends on a short list of facts about your site, and regulators have written down the tests. This post gives you those tests, the sources, and a way to run them on your own WordPress site. It belongs to our privacy and browsers series; the hub is Consent Mode v2 on WordPress.
This is not legal advice. I summarise published regulator texts and link them so you can read them yourself. Pages in this area get revised, so check the date on each one. Where the stakes are real, ask a privacy lawyer or your data protection officer. Facts below were checked in October 2026.
A banner is a tool, not a legal basis
A cookie banner is a user interface. It collects a choice, and it can be configured well or badly. It does not by itself make anything lawful, and the absence of one does not make anything unlawful. What matters is whether the law required consent for what your site does, and, if it did, whether you got a valid one.
Two laws are in play, and they are easy to mix up. I go through them in detail in Does server-side tracking bypass consent?, so here is the short version:
- ePrivacy Directive, Article 5(3): storing information on a device, or reading information already stored there, needs consent unless an exemption applies. This is what “cookie law” refers to. It applies to information in general, not only personal data.
- GDPR: any processing of personal data needs a lawful basis such as consent or legitimate interests. Counting visits with an IP address in a log is processing personal data.
A banner answers question 1. Question 2 needs its own answer even on a site with no banner at all.
What Article 5(3) exempts, and why analytics is not obviously in
Article 5(3) allows storage or access without consent in two cases: for the sole purpose of carrying out the transmission of a communication, or where it is strictly necessary to provide a service the user explicitly requested (text of Article 5 on legislation.gov.uk, a UK government copy of the EU directive; the EU’s consolidated version is on EUR-Lex).
The Article 29 Working Party, the EDPB’s predecessor, looked at analytics specifically in its Opinion 04/2012 (adopted 7 June 2012, section 4.3). Its reasoning: analytics is often considered strictly necessary by website operators, but it is not strictly necessary to provide a function the user asked for, because the user can use every feature with those cookies disabled. So first-party analytics cookies fall outside the two exemptions. The opinion then adds that they are “not likely to create a privacy risk” when strictly limited to first-party aggregated statistics, with clear information in the privacy policy, a user-friendly opt-out, and anonymisation of identifiable data such as IP addresses. It suggested the legislator might add a third exemption for first-party anonymised, aggregated statistics. The Article 5(3) text linked above contains no such third exemption, so any exemption in practice comes from national law and authority guidance.
The EDPB’s Guidelines 2/2023 (final version adopted 7 October 2024) do not settle the analytics question either. They say the exemptions are outside their scope (paragraph 4) and that Article 5(3) applying does not always mean consent is needed; each case is assessed for consent or an exemption (paragraphs 40 and 56). What they do widen is the scope: cookies, API calls, and JavaScript instructing a browser to send information back are all within Article 5(3) (paragraphs 32 and 33), and so are tracking pixels (paragraph 51). That has a practical consequence for the next section.
What national authorities say
This is the part that varies by country. The table summarises the sources I could read first-hand; the paragraphs below give the details.
France: the CNIL’s audience-measurement exemption
France is the clearest case of an exemption written down as a checklist. The CNIL’s page on audience measurement tools (dated 4 July 2025, French; my translation) says trackers can be exempt from consent if they meet these conditions:
- Purpose: strictly limited to measuring the audience of the site or app, on the publisher’s own account.
- Output: anonymous statistical data only.
- No cross-referencing or onward sharing: the data must not be combined with other processing, and non-anonymous data must not be passed to third parties.
- No cross-site tracking: the same identifier must not be used across several sites or apps to cross-reference, deduplicate or measure combined reach.
The page also gives recommendations, not just conditions: inform users (for example in the privacy policy), keep the tracker’s lifetime to 13 months without automatic extension, keep collected data for at most 25 months, and review these periods regularly. It adds that providers who say they reuse the data for their own account take the tool outside the exemption, and that some commercial offerings let you switch that reuse off through configuration.
Read that last point twice. The exemption belongs to a specific configuration of a specific tool, not to a product name.
Denmark: Datatilsynet and the Digitaliseringsstyrelsen
The Danish joint guidance (May 2025, Danish; my translation, section 3.1) says use of cookies and similar technologies may, as a starting point, only happen after consent, and gives visit statistics and personalised ads as examples where you must collect it. Cookies and similar technologies can be used without consent only where they are technically necessary for the service to work as intended and the information is not used for other purposes. Its examples are login, shopping baskets, user-chosen settings, authentication and load balancing. Analytics is not among them. The guidance also says you stay responsible even if a partner runs the technology (section 3.6).
Ireland: the DPC
The Irish Data Protection Commission’s guidance note (version last updated April 2020) asks “Do analytics cookies require consent?” and answers “Yes.” It then repeats the Working Party’s view that first-party analytics cookies are not likely to create a privacy risk when strictly limited to aggregated statistics, with clear information and an opt-out, and adds that it is unlikely such cookies would be a priority for enforcement action. That is a statement about enforcement priorities, not an exemption. It is also the oldest source here, so check whether the DPC has issued anything newer.
Germany: the DSK declines a blanket answer
The German authorities’ joint Orientierungshilfe for digital services (version 1.2, November 2024, German; my translation) deliberately does not say whether reach measurement may generally be done without consent (paragraphs 87 to 91). Its reasons: the term is too vague; modern tools process far more than a count of page views; and the purpose matters. A determination can at most be made for a precisely defined configuration and purpose, and stops being valid if further user information or another evaluation result is added. It describes a pure count of page views as something a simple counter or a log file without personal data can do, and contrasts that with third-party reach services that process extensive information about users. Even a simple visitor count is not automatically part of the service the visitor asked for; it depends on the purpose.
So in Germany “is analytics exempt?” has no one-line answer from the authorities, which is an answer in itself.
Common setups, tested against those sources
I will not tell you that any product needs no banner. What I can do is show where each kind of setup lands when you hold it against the conditions above. Treat these as prompts for your own check.
GA4 with default settings
GA4 sets two first-party cookies by default. According to Google’s cookie documentation, _ga distinguishes users and _ga_<container-id> persists session state, each with a 2 year expiry. That is longer than the 13 months the CNIL recommends, and the Danish and Irish guidance start from consent for analytics. In practice, a standard GA4 install is treated as consent-requiring in the jurisdictions above unless you can show it meets a local exemption, which for GA4 means checking the provider-reuse point in the CNIL text and the contractual terms yourself.
GA4 with Consent Mode
Consent Mode is a signalling mechanism: your banner tells Google tags what the visitor chose. Google’s documentation says that when analytics_storage is denied, tags do not read or write first-party analytics cookies, but in advanced mode cookieless pings are still sent, and those pings can include user agent, screen resolution and IP address. Google adds that privacy laws are region-specific and suggests configuring defaults by region. Whether such pings themselves need consent is a legal question Google’s page does not answer for you; the EDPB’s broad reading of “gaining access” is a reason to ask your authority, not to assume. Our Consent Mode v2 guide covers the mechanics.
Cookieless, first-party measurement
Some tools avoid cookies and local storage and count visitors another way. This removes one trigger but not the others. Ask three things. Does a script still read device data and send it back (EDPB Guidelines 2/2023 treat that as in scope)? Does the output stay anonymous in the CNIL’s sense, with no cross-referencing? Does the provider reuse the data for its own account? “Cookieless” in a marketing page is a starting point for the audit, not its result.
Server-log analytics
Counting from web server logs reads nothing from the visitor’s device, so Article 5(3) is not engaged by that counting. The DSK paragraph above describes a count based on log files without personal data as the simple end of the spectrum. But log files normally contain IP addresses, so the GDPR still applies: you need a lawful basis, a retention limit and information in your privacy policy. The trade-off is depth: logs cannot see scroll depth, events or conversions, and they count bots and cached requests differently. Our post on why WordPress analytics data goes missing explains how the gaps open up.
Server-side forwarding to ad platforms
Sending events to Meta or Google Ads is a different purpose from measuring your own audience. It involves a third party using the data for its own purposes, which the CNIL’s conditions rule out of the exemption. That topic has its own post: Does server-side tracking bypass consent?
Audit what your WordPress site actually sets
You cannot apply any of this until you know what your site stores. WordPress sites are good at surprises: the theme, a form plugin, an embedded video, a chat widget and an ads tag can each add their own storage and third-party requests. Run this audit on a staging copy or the live site.
- Open a clean private window without extensions and visit the home page. Do not click the banner.
- Open DevTools (F12) and go to the Application tab in Chrome or Edge, or Storage in Firefox. Note every entry under Cookies, Local storage and Session storage: name, domain, expiry.
- Open the Network tab, reload, and note requests to other domains. A pixel or script that sends device data back can matter even without a cookie.
- Clear site data, reload and click Reject. Compare. Anything non-essential still present is your first fix.
- Clear again and click Accept. The new entries are what your banner is gating. Record their lifetimes and compare them with the numbers in your authority’s guidance.
- Write one row per item: name, who sets it, purpose, lifetime, and which branch of the decision tree it lands on. Repeat after plugin, theme or tag changes.
Cookie scanners can speed this up, with a caveat from the regulators themselves. The EDPB’s Cookie Banner Taskforce noted in its report (adopted 17 January 2023, paragraph 29) that such tools list the cookies placed but do not establish their nature, so website owners must still document what each one is for. It also noted that cookie features change regularly (paragraph 27), which is why a one-off audit is not enough.
If you need a banner: what regulators stress
Choosing a plugin comes second. We list consent management options in the Consent Mode v2 post; any of them can be set up well or badly. These are the design points authorities have put in writing:
- Reject on the first layer. The EDPB taskforce report (paragraph 8) says a vast majority of authorities consider the absence of a refuse option on any layer with a consent button to be an infringement; a few disagreed.
- No pre-ticked boxes. The taskforce confirmed they do not give valid consent (paragraph 10), citing GDPR recital 32.
- No hidden reject link. A reject that is only a link in body text or outside the banner, without visual support, is not valid consent on the taskforce’s examples (paragraphs 11 to 14).
- Equal prominence. The DPC says that if you offer an accept button you must give equal prominence to a reject option, or one that leads to managing cookies. The Danish guidance shows a green accept button next to a barely visible reject button as an example of nudging. The taskforce is more cautious on colours: it says no general standard can be imposed and judges case by case, but illegible reject text is manifestly misleading (paragraphs 17 and 18).
- Legitimate interests cannot be the basis for Article 5(3) storage. The taskforce confirmed this (paragraph 24).
The taskforce report itself says its positions are a minimum threshold, not a green light from any authority (disclaimer, page 3). National rules still apply on top.
Three worked examples
A brochure site in France with a log-based counter. Nothing is stored on the device, so Article 5(3) is not engaged by the counter; the GDPR still needs a basis for log IP addresses. On that reading the counter itself does not trigger Article 5(3). Any embedded video or chat widget would need its own audit.
A WooCommerce store in Denmark with standard GA4. The Danish guidance lists visit statistics as a consent example. GA4’s cookies are not on the technically-necessary list. Plan for consent before the tag fires, a visible reject, and a test that nothing loads after refusal.
A blog in Germany with a cookieless tool. The DSK will not say that reach measurement is generally exempt. Document the exact configuration, what data is collected, whether the provider reuses it, and ask your supervisory authority or counsel. If unsure, treat it as consent-requiring until you have an answer.
Where Traculo fits
Traculo is a web app with a WordPress plugin for website analytics, goals and funnels, and server-side event forwarding, and it includes a built-in data processing agreement. I am deliberately not telling you whether it, or any other analytics tool, needs a banner on your site. That depends on your configuration, your country and what you forward where. What I can state is how the plugin behaves: its browser script sets no cookies or local storage, but its server sets one _ga-compatible cookie (HttpOnly, two years) unless consent is denied through the WP Consent API (statistics category) or a filter. It ships no banner and does not hold back collection itself, so it is not a cookieless tool. Run the audit above on it as you would on anything else, and read the data processing agreement alongside your own records.
Frequently asked questions
Does first-party analytics need a cookie banner?
The Working Party said in 2012 that first-party analytics cookies are not exempt but low-risk with safeguards, and Ireland’s DPC says analytics cookies require consent. France’s CNIL lists conditions under which audience measurement can be exempt. Denmark starts from consent. The answer depends on your authority and your setup.
Is “cookieless” analytics automatically exempt?
No. Article 5(3) also covers reading information from the device and scripts or pixels that send device data back (EDPB Guidelines 2/2023), and the GDPR applies to any personal data, such as IP addresses. Test the tool against your authority’s conditions.
What are the CNIL’s conditions?
Measure the site’s own audience for the publisher only, produce anonymous statistics only, no cross-referencing or non-anonymous data to third parties, and no tracking of the same person across sites or apps. It recommends a 13 month tracker lifetime and 25 month data retention. See the CNIL page.
Is a consent plugin enough?
Only if it is configured to block what needs consent until it is given, offers an equally visible reject, and you have checked the result in DevTools. A plugin can display a banner while tags still fire behind it.
Related reading
- Consent Mode v2 on WordPress: the four signals
- Does server-side tracking bypass consent?
- GDPR and WordPress analytics: the Danish picture
- Safari ITP and WordPress: what the 7-day cookie cap really affects
- How much data do ad blockers cost a WordPress site?
- Best GA4 alternatives for WordPress
- Client-side vs server-side tracking in WordPress
Want to see your own traffic without guessing at the numbers? Look at Traculo website analytics, then run the audit above on whichever tool you choose.