Report · 2027 edition

State of UK Tracking 2027

Marc Alexander's automated testing, Tracking Watch, visited 131 UK websites in a real browser from a UK connection, between 16 November and 21 November 2026. It accepted, rejected and ignored their cookie banners, typed a test email into their site search, and read 34,401 tracking requests. Most sites get consent wrong somewhere.

  • 63%track visitors before they choose
  • 56%keep tracking after "Reject all"
  • 43%pass an email typed into search to a tracking tool
  • 76%have at least one consent problem

Share a finding: 63% track visitors before they choose · 56% keep tracking after “Reject all” · 43% pass a search email to a tracking tool · 76% have at least one consent problem

What the tests found.

The share of the 131 sites with each problem. Choose a sector to compare it with every site, shown by the thin line.

  1. Track visitors before they choose 63% 82 of 131
  2. Keep tracking after “Reject all” 56% 74 of 131
  3. No “Reject all” on the banner’s first screen 34% 45 of 131
  4. Optional cookies switched on in the banner’s settings 13% 17 of 131
  5. Send an email typed into site search to a tracking tool 43% 56 of 131
  6. Consent Mode v2 missing or wrong 25% 33 of 131
  7. Count GA4 page views twice 13% 17 of 131
  8. Broken pages in their own sitemap 24% 31 of 131
  9. Old Universal Analytics tags still in Tag Manager estimate 38% 50 of 131
  10. Lose every GA4 event to a common ad blocker estimate 91% 119 of 131
  11. Returning Safari visitors look new after a week estimate 83% 109 of 131

Each problem was seen in the browser and saved with the request that proves it. "Estimate" marks the three that rest on the scan's measurement of cookie lifetimes and blocked requests, rather than a request seen on the site.

Of the sites that pass a search email on, 51 send it to an ad platform such as Google Ads or Meta.

By sector.

The share of sites in each sector with the problem. Darker cells mean more of the sector is affected. Small sectors move a lot with one site, so read them as a guide.

Sector (sites)Track visitors before they chooseKeep tracking after “Reject all”No “Reject all” on the banner’s first screenOptional cookies switched on in the banner’s settingsSend an email typed into site search to a tracking toolConsent Mode v2 missing or wrongCount GA4 page views twiceBroken pages in their own sitemap
Retail (31)68%58%35%16%68%19%19%29%
Finance and payments (19)53%42%37%11%37%42%0%16%
Lead generation (18)56%50%50%22%33%22%11%28%
Travel (15)53%67%7%20%47%27%0%20%
Property (14)71%64%36%7%21%36%0%21%
Subscriptions (11)73%64%36%9%9%9%36%36%
Charities (10)70%70%40%0%60%10%10%20%
Health (8)63%50%38%13%25%50%25%25%
Tourism boards (5)60%40%20%0%60%0%40%0%

What it means.

Consent is where most sites fail

A banner that's switched on isn't proof that it works. On 56% of sites, the tags carried on after the visitor pressed "Reject all", usually because one tag in Tag Manager never checked for consent.

Search terms leak personal data

Visitors type email addresses and order numbers into search boxes. When the results page puts the term in its address, every tag sends it on, so 56 of the 131 sites handed a test email to GA4 or an ad platform.

GA4 shows less than really happens

A common ad blocker stopped GA4 completely on 91% of sites, and Safari's cookie limits make returning visitors look new on 83%. Neither is a fault, but both matter before you trust a report.

How to check your own site.

The free tracking checker shows what your homepage sends before a choice.

For the full set of tests in this report, with the fix for each problem, the GA4 mini audit is from £175. Fixing consent itself is covered by Consent Mode v2 implementation.

How the tests worked.

Each site was visited by Tracking Watch, Marc's automated testing, on desktop, phone and tablet: once without answering the banner, once after accepting and once after "Reject all", plus a sample of up to 40 pages from its sitemap. It used no logins, placed no orders and sent no forms.

Only the latest scan of each site counts, and scans stopped early by a bot check or a region redirect were left out. Earlier results for online retailers and property websites are in the retail scan and the property scan.

Questions about the report.

How were the websites chosen?

They are UK organisations of the kind Marc has worked for: online shops, lead-generation firms, property, finance and payments, health, travel, tourism boards, charities and subscription businesses. A site was included if it could be reached from the UK without a bot check and loaded some tracking. Nobody was picked because of a known problem.

Why is it called 2027 when the tests ran in 2026?

It's the edition for the year ahead. Every scan ran between 16 November and 21 November 2026, and the numbers describe the sites as they were then.

Can I find out which sites were tested?

No. The report is anonymised, and the companies haven't been told their results in public. Marc can run the same tests on your own site, with a written report.

Is tracking before a choice illegal in the UK?

Non-essential cookies and similar tracking need consent under PECR, and the ICO's guidance says that consent has to come before the tracking starts. The GA4 and UK GDPR guide explains what that means for GA4.

Find out where your site stands.

Tell Marc which site you'd like tested. The GA4 mini audit runs the same tests, with a written report and the fix for each problem.

Mini audit from £175.

Prefer to talk first? Choose "20-minute intro call" in the form.
Or email [email protected].

Marc replies personally, usually within one working day.