The State of Conversion Tracking in 2026: What the Sourced Data Shows

What is the state of conversion tracking in 2026?

Pixels still see most conversions, but several independent mechanisms each hide some, and no published source measures the total. What can be measured is the pieces: about three in ten internet users run an ad blocker, about one in five page views comes from Safari or Firefox, and platforms report server-side gains of 8% to 19% on their own measures.

This page summarizes the sourced figures from our three research reports and shows how to combine them. Every figure links to its publication in Sources and methodology, and the date at the top shows when we last re-checked it.

Measured figures at a glance

What does no source measure?

No source we could verify gives the share of a typical store's conversions that are missed once ad blockers, browser rules, consent choices and checkout redirects are combined. Ranges such as "30-45% of conversions invisible" or "95-99% capture with server-side tracking" appear in many articles, but we could not trace them to a primary publication. Earlier versions of this report included some of them, and they are listed under Corrections and updates.

Which browsers limit tracking in 2026?

Safari blocks all third-party cookies, Firefox confines each one to the site that created it, Edge blocks storage for classified advertising trackers, and Chrome allows them by default. Safari and Firefox together supplied 21.0% of page views worldwide and 33.7% in the United States in September 2026.

Safari also limits cookies that a site's own scripts write. WebKit documents that it deletes them after 7 days without interaction and caps them at 1 day when a tracker-classified site sends the visitor to a URL with a query string or fragment. Those rules reach every iOS browser outside the EU and Japan, because Apple requires them to use WebKit.

Google kept third-party cookies in Chrome and retired most of the Privacy Sandbox APIs in October 2025. The share of Safari and Firefox rose from 16.0% to 21.0% worldwide between September 2025 and September 2026, but StatCounter is a sample and year-over-year changes can reflect detection changes. Details, the exposure calculation and the limitations are in our third-party cookie statistics.

How many people use ad blockers in 2026?

An estimated 1.1 billion people used ad blockers in Q4 2025 according to eyeo, and 29.5% of internet users aged 16 and over said they use one at least sometimes in GWI's Q2 2025 survey. The count is rising while the share is falling, because the internet population grew faster.

The two headline numbers count different things. On a base of 6.04 billion people online, 1.1 billion users is 18.2%, not 30%, and we could not reconcile eyeo's statement that its cohort represents 30% of the internet population with its own user count. Cite the count and the rate separately. Country rates, filter-list coverage and the CPA effect are in our ad blocker statistics.

How much does server-side tracking add?

Platforms report gains of 8% to 19% on their own measures, and the one independent test we found averaged +16% tracked conversions with a wide spread. Each platform measures something different, so the figures cannot be averaged into a single recovery rate.

These are company claims or small tests, not audits. The method behind each, and a vendor case study that we recalculated, are in our benchmark report.

How do uplift and loss relate?

Uplift is measured against what the pixel saw, and loss is measured against what actually happened. If server-side delivery adds u percent more conversions, the pixel missed u divided by (1 + u) of the total.

How much ad traffic is invalid?

Pixalate measured invalid traffic of 23% for web, 36% for mobile apps and 21% for connected TV in global open programmatic advertising in Q4 2025. Those rates describe programmatic inventory, not paid search or paid social, so they do not tell you how many of your Google Ads or Meta clicks are invalid.

We found no primary published invalid click rate for Google Ads or Meta. Bot traffic can inflate conversion counts instead of removing them, so the effect on tracking runs in the opposite direction from ad blockers. Compare clicks with sessions and filter known bots before events reach the ad platforms. Our bot traffic statistics show the sourced bot shares and the formulas for how bots distort conversion rate and CPA.

How do the layers of loss compound?

When several mechanisms each hide a share of conversions, the visible share is the product of what each leaves, not the sum of what each removes. Three layers that hide 10%, 8% and 5% leave 78.7% visible, so 21.3% is missing and server-side delivery would need to add about 27% to restore the total.

The inputs here are examples, not measurements, and the formula assumes the layers are independent. In practice they overlap, for instance when one visitor has both an ad blocker and Safari, so the true visible share is usually higher than the product suggests. Replace each input with a number from your own data.

If blockers hide a share h of your conversions from the pixel, your reported CPA is the real CPA divided by (1 - h). At 30% hidden, the reported CPA looks 42.9% higher than the real one.

How can you measure your own conversion gap?

Compare orders in your store backend with the conversions each ad platform reports for the same period, then split the difference by browser and device. That comparison is the only measure of loss that applies to your audience.

Attribution models differ, so expect a gap even with perfect tracking. Look for a gap that is larger in Safari than in Chrome, or one that grows after a browser update. For setup, see our guides on server-side tracking and Event Match Quality, and our page on server-side tracking.

Why do other 2026 tracking reports show different numbers?

They often mix denominators, combine surveys with measurements, or repeat figures that were never sourced. Three patterns explain most differences.

A fourth pattern is ranges with no origin, such as "20-40% of conversions recovered" or "95-99% total capture". We searched for a primary source for each and found none, so this report does not repeat them.

Frequently Asked Questions

What percentage of conversions does pixel-only tracking miss?

No source we could verify measures it for a typical store. The measured components are 29.5% of internet users using an ad blocker, Safari's 7-day cookie cap and 33.7% of US page views from Safari or Firefox. Compare your backend orders with platform conversions to find your own figure.

Did Chrome remove third-party cookies?

No. Chrome still allows third-party cookies by default, and Google said in April 2025 that it would not add a standalone prompt. Safari blocks them all, and Firefox partitions them per site.

How many people use ad blockers?

eyeo estimates 1.1 billion active ad-blocking users in Q4 2025, and GWI's survey found 29.5% of internet users aged 16 and over use one at least sometimes. The two figures count different things and do not reconcile, so cite them separately.

Is server-side tracking worth it?

Published gains range from 8% to 19% on platform-reported measures and +16% on average in one independent test, so the value depends on how many conversions your pixel misses today. Measure that gap first, then decide.

Is Event Match Quality linked to lower CPA?

Meta publishes no figure that links Event Match Quality scores to cost per result. Meta says high-quality matching may improve attribution and performance, and it offers Additional Conversions Reported as the account-level measure.

Sources and methodology

Primary sources

• eyeo, 2026 Ad-Blocking Report (May 2026): user counts, the 2021 and 2025 rates and the definition of an ad-blocking user.
• We Are Social, Meltwater and Kepios, Digital 2026 Global Overview Report (October 2025 edition): GWI ad-blocker data and 6.04 billion internet users.
• StatCounter Global Stats, browser market share (worldwide, September 2026): worldwide shares, plus the September 2025 comparison month.
• StatCounter, browser market share in the United States: US shares.
• WebKit, Tracking Prevention in WebKit: third-party cookie blocking and the 7-day and 1-day caps.
• Apple, App Review Guidelines (2.5.6): browsers must use WebKit, with entitlements for alternative engines in the EU and Japan.
• Microsoft, Tracking prevention in Microsoft Edge: the Balanced default.
• Mozilla, Firefox rolls out Total Cookie Protection by default (June 14, 2022): Firefox behaviour.
• Google, Next steps for Privacy Sandbox and tracking protections in Chrome (April 22, 2025) and Update on plans for Privacy Sandbox technologies (October 17, 2025): Chrome's cookie policy and the retired APIs.
• Adjust, Mobile app trends: 2026 edition: App Tracking Transparency opt-in rates.
• Meta for Business, Conversions API: the 17.8% lower cost per result figure.
• Meta for Developers, Dataset Quality API: Event Match Quality, Additional Conversions Reported and event coverage.
• TikTok for Business, Performance Fundamentals: 19% more events and a 15% better CPA.
• Google, Enhanced conversions: 8% incremental ROAS from 99 conversion lift studies.
• Workshop Digital, Benefits of Enhanced Conversions in Google Ads: the five-account test.
• Pixalate, Q4 2025 North America Invalid Traffic (IVT) and Ad Fraud Benchmarks (March 9, 2026): global, US and Canada rates.

How we verified the figures

• This page summarizes figures that were each read from the original publication for our ad blocker, third-party cookie and benchmark reports. Those pages show the page numbers and calculations.
• Calculated values use only the inputs shown: uplift loss is u divided by (1 + u), mobile share is 629 divided by 1,077, and the layered example multiplies the share still visible at each layer.
• Evidence types are our labels: platform-reported means the company that sells the ads published it, vendor means a company that sells related tools did, and independent means neither.

Limitations

• Several sources have an interest in the topic. eyeo makes Adblock Plus, Adjust and Pixalate sell measurement and fraud tools, and the ad platforms sell the advertising.
• StatCounter is a sample of page views from sites that use its code, not a census of web traffic.
• The Workshop Digital test covers five accounts and Google Ads leads, so it does not generalize.
• The layered-loss example uses assumed inputs and assumes independence.

Corrections and updates

• October 7, 2026: Rewritten from the version published September 15, 2026. We removed figures we could not trace to a primary source or that disagreed with the one we found: Chrome 71.4%, Safari 14.8%, Firefox 2.3% and Edge 4.7% of page views (StatCounter's September 2026 figures are 66.45%, 18.20%, 2.79% and 6.02%); "17-20%" of web traffic as cookieless (measured: 21.0% for Safari and Firefox); server-side adoption of 12% rising to 34% of e-commerce sites; a 30-45% pixel-only conversion gap and the pixel coverage rates by audience segment; 15-25% ATT opt-in and 75-85% opt-out (Adjust measured 38% opt-in); 95-99% total conversion capture; 20-40% more attributed conversions and 13-25% lower CPAs within 30 days; the affiliate server-side adoption trajectory; the 47.2% bot share of web traffic; and the 37% desktop ad-blocker rate and revenue example based on it.