What the Numbers Say When Two Analytics Tools Disagree
A recent post from Plausible claimed that 58% of tech-savvy audiences block Google Analytics. The figure came from a three-day window where the author saw Plausible report 50.9k unique visitors while Google Analytics logged just 21.1k. That's a massive gap, and it made me wonder whether my own Google Analytics numbers are undercounting my site's traffic. I decided to run both tools side by side for a week to find out.
Here's what Google Analytics reported for the same period:
Side by side:
| Metric | Plausible | Google Analytics |
|---|---|---|
| Unique Visitors | 973k | 841k |
| Pageviews | 1.4m | 1.5m |
| Bounce Rate | 82% | 82% |
| Visit Duration | 1m 31s | 1m 24s |
The results are not straightforward. Plausible shows 15% more unique visitors than Google Analytics, which is notable but nowhere near the doubling suggested by the viral blog post. However, Plausible reports 5% less raw traffic — the metric that matters most for an ad-supported site. Neither tool gives a definitive answer about which number is closer to the truth.
Why the Two Tools Disagree
Plausible is still third-party JavaScript, just like Google Analytics. It's less known and not Google, so fewer people likely block it, but it remains just as blockable. It is not a fundamentally different measurement method.
Server-side analytics, such as Netlify Analytics, use server logs that cannot be blocked by client-side ad blockers. That approach brings its own complications, including counting bots that Google Analytics filters out. The alternative of implementing server-side Google Analytics is technically sound but adds significant infrastructure overhead compared to dropping a script tag into a page.
The Case for Sending Data Once
Running multiple client-side analytics scripts simultaneously felt redundant. For sites tracking custom events, duplicating that logic across tools becomes especially messy. An abstraction layer like Analytics can help manage multiple destinations, but a cleaner approach is to send data from the client once and handle distribution server-side, similar to what Segment does — though that can get expensive. Self-hosting or building a simple internal relay is feasible. Reducing the amount of third-party data crossing the wire and relying on first-party JavaScript seems like a sensible pattern going forward.
