CloudFlare give me two different measures of how many people have visited my website, one using their Analytics product and other from their new Web Analytics product. I get very different results for page views and number of visitors. I also get a third distinct set of results from my own analytics tool.
Most of my visitors read the technical articles and therefore the audience is probably very technical and likely to be using an ad blocker. If this is the case then one way to test my hypothesis would be to write some articles that appeal to a non-technical audience who are less likely to use an ad blocker. In this case I would expect the two analytics methods to agree more closely.
CloudFlare's (normal) analytics say I've had 234 unique visitors. But their Web Analytics tool says I've had 11 visitors. My own tool reports 12 unique visitors.
Why are these results so different? Maybe one measure might be including bots and another might only be trying to report real people using normal browsers, but the difference seems too high for that1. I'd also expect real usage to fall when its night in the countries I get most traffic from, which I don't see. Perhaps the difference is caused by 95% of my readers are using an ad blocker.
My own analytics tool can't give results from a rolling 24 hour window, it only groups data by day. Therefore I recorded the values at 11pm, which should be close enough. My simple method of logging IP addresses when a page is loaded and counting the unique IP addresses each day says that I've had 12 unique users. Much closer to the CloudFlare analytics beta result, but I wouldn't expect my bespoke tool to be blocked by an Ad Blocker.
If it were as simple as concluding that my own results agree with the CloudFlare analytics beta then that might be enough. But they only agree on this particular metric. I've logged 47 page views today using my own tool but the CloudFlare Analytics beta reports only 11 page views2.
Please let me know on twitter if you have any ideas!