Tony Wang6 min readAnalytics Vendors Split by Sophistication, Not by Privacy
Analytics tools split top-heavy vs. tail-heavy by product sophistication, not privacy - and a third of top sites show none at all.
Split this census's analytics tags by rank tier expecting the same "paid/deliberate vs. free/default" story that's run through nearly every other category in this series, and the privacy-focused tools — the ones you'd guess a deliberate, values-driven team would pick — don't cooperate. The real split runs along a completely different line.
Privacy-focused analytics don't skew the way you'd expect
| Vendor | Top 10K | 10K-100K | 100K-1M |
|---|---|---|---|
| Amplitude5.78x top-heavy | 0.67% | 0.411% | 0.116% |
| Mixpanel1.6x top-heavy | 0.15% | 0.222% | 0.094% |
| Matomoroughly flat | 1.3% | 1.351% | 1.064% |
| Google Analyticsmildly tail-heavy | 17.51% | 21.646% | 21.625% |
| Meta Pixelmildly tail-heavy | 4.15% | 6.2% | 7.184% |
| Plausibleno clear skew | 0.16% | 0.31% | 0.188% |
| Fathomno clear skew | 0.08% | 0.097% | 0.082% |
The intuitive guess going in was that Plausible, Matomo, Fathom, and PostHog — all positioned as privacy-conscious alternatives to Google Analytics, all requiring a team to actively seek out and choose them over the free default — would skew top-heavy the same way Google Workspace, DigitalOcean's competitor Proton Mail, and every other "deliberate choice" finding in this series has. They don't. Matomo comes closest to flat (1.30% → 1.06%), and Plausible and Fathom show no consistent direction at all across the three bands. Privacy-consciousness alone, it turns out, isn't the axis that predicts where a technology concentrates on the web.
The real split is product-analytics sophistication
Amplitude and Mixpanel aren't privacy tools — they're product analytics platforms, built for tracking detailed user behavior, funnels, and feature adoption rather than counting pageviews. They cost real money at scale and require a team whose job is specifically to study user behavior, not just marketing. Amplitude's 5.78× gradient is sharper than almost every other rank-tier finding in this census series. A spot-check of its top-10,000 domains reads exactly like what that number predicts: digitalocean.com, unity.com (the game engine company), msnbc.com, decathlon.in, infobip.com, yotpo.com — companies with the kind of dedicated product or growth function that would deliberately invest in this category of tool, not sites that stumbled into it.
Google Analytics and Meta Pixel — the two most widely deployed tags in this census by a wide margin — sit at the opposite end, both mildly tail-heavy. Neither requires a deliberate decision the way Amplitude does; they're close to the default expectation for "a site that tracks anything at all."
A third of even the most-visited sites show no analytics at all
This is a separate finding from the vendor breakdown above, and the flatness itself is the surprise. Roughly a third to two-fifths of reachable sites in every single rank band — including the top 10,000, the most-visited sites on the entire web by this census's own measure — carry no detected analytics tag at all. Before assuming that's just search engines, CDNs, and cloud consoles skewing the top-tier number, we checked: only 2.8% of the zero-analytics top-10,000 group is flagged as infrastructure. The other 97% are genuine, human-facing sites that this static-HTML fingerprint simply can't see any analytics on.
The likely explanation isn't "these sites don't measure anything" — it's that a growing share of analytics happens somewhere this detection method can't reach: server-side tag managers, first-party data pipelines, or custom-built internal analytics that never touch the homepage's client-side JavaScript. As privacy regulation pushes measurement further server-side, a homepage-HTML fingerprint is likely to see less of it over time, not more — worth tracking in future refreshes of this census rather than treating as a one-time number.
Limits, stated plainly
Amplitude and Mixpanel's counts, while large enough to trust the direction, are still a small fraction of the census overall (a few thousand sites each) — read the exact ratios as directional. "No detected analytics" is a statement about detectability, not about measurement: a site using server-side analytics, a custom in-house pipeline, or a vendor this signature set doesn't yet recognize would also show up in that 35-40%, indistinguishable from a site tracking nothing at all. As with the rest of this census, this is a homepage-only, static-HTML fingerprint — it can't see what loads after a click, a login, or a page navigation deeper into the site.
Query the full Tech Stack dataset
1,000,000 fingerprinted domains — CDN, CMS, ecommerce, analytics and more — over one REST API, ranked by Tranco traffic rank.
Frequently asked questions
Do privacy-focused analytics tools skew toward high-traffic websites?
No - a genuine null result. In Crawlora's 1,000,000-domain tech-stack census, privacy-focused analytics tools (Plausible, Matomo, Fathom) show no consistent rank-tier skew in either direction, contrary to the expectation that a deliberately-chosen privacy tool would concentrate among high-traffic sites the way other deliberate-choice technologies do.
Which analytics tool shows the strongest correlation with website traffic rank?
Amplitude, a premium product-analytics platform. It runs 5.78x more often among the top 10,000 Tranco-ranked sites than in the 100K-1M tail - one of the sharpest rank-tier gradients found anywhere in this census series. Mixpanel shows the same direction more mildly (1.6x).
Which companies use Amplitude for analytics?
A spot-check of Amplitude's top-10,000 domains in this census found digitalocean.com, unity.com (the game engine company), msnbc.com, decathlon.in, infobip.com, and yotpo.com - companies with the kind of dedicated product or growth function that would deliberately invest in detailed behavioral analytics.
What percentage of top websites have no detected analytics tag?
37.78% of the top 10,000 Tranco-ranked sites show no detected analytics-category technology at all - nearly identical to the 34.46% and 39.44% rates in the 10K-100K and 100K-1M bands. This flat rate isn't explained by infrastructure domains, which make up only 2.8% of the zero-analytics top-10,000 group; the rest are genuine, human-facing sites whose analytics (likely server-side or first-party) simply aren't visible to a homepage-HTML fingerprint.