real numbers
Cloudflare vs Google Analytics: Our Real Traffic Numbers for 28 Days
What is the direct answer?
For the same 28 days on prismpublication.com, Cloudflare's server-side analytics counted 34,149 page views. Google Analytics 4 counted 545. Cloudflare's own browser-based Web Analytics estimated about 500. The two browser tools agree with each other; the server count is about 63 times higher because it counts every page the server sends, including to bots.
What are the key takeaways?
- Cloudflare edge: 34,149 page views and 78,754 requests in 28 days.
- Google Analytics 4: 545 page views, 147 sessions, 118 active users.
- Cloudflare Web Analytics (browser beacon): about 500 page loads and about 320 visits, estimated from a sample.
- When a real 404 page replaced a catch-all homepage on September 27, Cloudflare's daily page views fell 83 percent within two days.
- Nothing here is rounded for effect. Small, honest numbers are the point.
Why publish numbers this small?
Because made-up statistics have become a growth tactic. People publish invented figures, wait for AI assistants to repeat them, and call it a win. This page goes the other way: a small site, real dashboards, the same date range in every tool, and the method written down so anyone can check it.
A number you can check is worth more than a big number you cannot. If you run a small site and your Cloudflare dashboard looks far healthier than Google Analytics, this is what that gap looks like from the inside.
What did each tool report for the same 28 days?
All figures cover September 10 to October 7, 2026, for prismpublication.com only. Google Analytics was filtered to the hostname prismpublication.com, because the same property also tracks two other sites.
| Metric | Cloudflare edge (server) | Cloudflare Web Analytics (browser) | Google Analytics 4 (browser) |
|---|---|---|---|
| Page views | 34,149 | about 500 (estimate) | 545 |
| Visits or sessions | not reported | about 320 (estimate) | 147 sessions |
| Visitors or users | about 340 unique visitors per day (daily counts, not de-duplicated across days) | not reported | 118 active users |
| All HTTP requests | 78,754 | not applicable | not applicable |
| Requests flagged as threats | 3,360 | not applicable | not applicable |
| Key events (leads) | not applicable | not applicable | 0 (no lead event is set up yet) |
The server count is about 63 times the Google Analytics count. The two browser counts are within 10 percent of each other. That pattern says the gap is about what each tool counts, not about one of them being broken.
What happened when a real 404 page went live?
On September 27, the site got a proper 404 page and about 45 redirects from old URLs. Before that, an address that did not exist still returned the homepage with a success code, which is called a soft 404. Bots that probe random addresses were getting a full page back, and Cloudflare counted each one as a page view.
| Daily average | Sep 10 to 26 (17 days) | Sep 27 to Oct 7 (11 days) | Change |
|---|---|---|---|
| Page views | 1,814 | 301 | down 83% |
| 404 responses | 78 | 1,656 | up about 21 times |
| All requests | 2,624 | 3,104 | up 18% |
| Unique visitors | 327 | 360 | up 10% |
The requests kept coming. They just stopped looking like page views. About 1,500 daily page views disappeared, and about 1,500 more 404 responses a day appeared in their place. That is what bots reading pages that do not exist look like. Google Analytics never counted them, because those bots do not run its tracking script. After the fix, Cloudflare's 301 page views a day are about 15 times Google Analytics' daily average of about 19, down from about 63 times.
Where did the Google Analytics visitors come from?
| Channel | Sessions | Engagement rate |
|---|---|---|
| Direct | 114 | 24.6% |
| Organic search | 13 | 69.2% |
| Referral | 8 | 62.5% |
| Organic social | 6 | 33.3% |
| Unassigned | 5 | 40.0% |
| AI assistant | 1 | 0% |
| Total | 147 | 31.3% |
Search visitors were few but engaged: 13 sessions with an average of more than four minutes each. One session came from an AI assistant. That is the honest starting point for a site that is trying to get cited by AI tools.
Which countries showed up, and what do they tell you?
| Country | Active users | Engagement rate | Average engagement time per user |
|---|---|---|---|
| United States | 75 | 7.9% | 2 seconds |
| Israel | 22 | 72.0% | 4 minutes 50 seconds |
| Not set | 7 | 0% | 0 seconds |
| India | 4 | 75.0% | 57 seconds |
| Ireland | 4 | 0% | 0 seconds |
Two honest flags. First, the Israel row includes my own visits. The filter that hides internal traffic is still in testing mode, so Google Analytics labels my visits but still counts them. Second, 75 United States users averaging two seconds look like automated or drive-by traffic that got past Google's bot filtering. Even the "clean" tool needs a second look before you quote it.
Which pages got the most views in Google Analytics?
| Page | Views |
|---|---|
| / (homepage) | 282 |
| /run-ads/ (now redirects to /ad-submission/) | 59 |
| /ad-submission/ | 22 |
| /demo/ | 21 |
| /blog/ | 17 |
| /developers/ | 12 |
How were the numbers collected?
- Date range: September 10 to October 7, 2026, read on October 8, 2026.
- Google Analytics 4: standard reports with the filter "Hostname contains prismpublication.com". Page views, top pages and the 252 figure below come from the Pages and screens report. Sessions and channels come from Traffic acquisition. Active users and countries come from Demographic details.
- Cloudflare edge: the zone's daily HTTP request totals for prismpublication.com, including page views, requests, unique visitors, response codes and threats.
- Cloudflare Web Analytics: the browser beacon for the host prismpublication.com. Cloudflare reports these from a sample (about 1 in 10 page loads in this range) and scales them up, so treat them as estimates.
- One open inconsistency: with the same filter, Google Analytics shows 118 active users in its user reports and 252 in its pages report. I report 118 as the user count and show the gap rather than hide it.
- Daily data: download the daily Cloudflare edge figures as CSV.
Why the tools count so differently is its own topic: why Cloudflare shows more visitors than Google Analytics.
Frequently asked questions
How big is the gap between Cloudflare and Google Analytics page views?
On prismpublication.com over 28 days, Cloudflare's edge counted 34,149 page views and Google Analytics 4 counted 545, about 63 times fewer. After a real 404 page went live on September 27, Cloudflare's daily page views were about 15 times Google Analytics' daily average.
Which number is closest to real human visitors?
The two tools that run in the browser agreed: Google Analytics 4 counted 545 page views and Cloudflare Web Analytics estimated about 500. Cloudflare's edge counts every request the server answers, including bots, so it was the outlier.
Are these numbers made up or rounded for effect?
No. They were read from the Google Analytics 4 property and the Cloudflare dashboard on October 8, 2026, for the same date range. The Cloudflare Web Analytics figures are Cloudflare's own estimates from a sample, and this page says so.
Did the site owner's own visits count?
Yes, in Google Analytics. The internal traffic filter was still in testing mode, so visits from Israel, where I work, are inflated. The Israel row is shown separately above.
AI-assisted, numbers pulled from the live dashboards and checked by a human.
Real questions, real matches
Prism shows a labeled ad in an AI chat only when it fits the question. Test your own ad against real buyer questions, free, and see the score for each one.