When is a page view not a page view?

Hi all -
Does anyone know if there’s update guidance since this blog post back in 2011 about how Cloudflare Analytics compares to “traditional” beacon based analytics systems like Google or Adobe? Is there any updated guidance from Cloudflare since this blog article was published?

I have a content director wanting to know how she can compare and validate that our Adobe Analytics implementation is ‘correct’. What filters can I put on CF analytics to make a comparison between the two and demonstrate the correlation between the two?

Thanks much,

Aaron

80% of statistics are fudged. The other 30% are misinterpreted.

What specific metric are you looking at?

If you really want to see what’s going on, get the raw data and crunch that. Logflare can help with that, but it’s going to cost money.

2 Likes

Haha! Too right. Thanks for the reply.

Trying to compare page views in Cloudflare to page views from Adobe Analytics. If I choose a specific URL, the stats are generally 10x higher in Cloudflare than Adobe. If I filter out anything with a ‘bot’ type of browser, that’s probably not firing javascript, that gets me closer to making a comparison, but I’m still at a loss of how else to compare.

We are starting to download the raw logs for this particular site, but ideally I don’t want to suggest to my business users that Cloudflare Analytics are what they should be utilizing for day-to-day site performance decisions.

As the article says, comparing actual logs vs beacon stats is quite difficult. I generally have my browsers set to block trackers, etc., so that’s going to throw things off as well. This Support article says the same, and is in line with that ten-year-old blog post.

If you mean for marketing, then I wouldn’t recommend it. I find Cloudflare’s analytics best for measuring how Cloudflare is handling my site: Firewall and Caching.

Agreed. Better filtering gets you closer. Looking at my own traffic, there’s just so much that comes from hosting ASNs (AWS/Google/OVH) that’s trying to disguise it as not a bot, but the ASN is pretty telling that it’s not a human visitor. If you can filter out those obvious ones, that will get you closer still.

Still, some sort of beacon, like Adobe, Google, etc., is probably your best bet, short of doing your own analytics from raw server or Cloudflare logs.

2 Likes

If you want to compare beacon-based analytics, then you should use Cloudflare Web Analytics, not the Domain Analytics - explanation here How accurate is cloudflare analytics? - #3 by freitasm

Also, read Cloudflare Web Analytics | Cloudflare for the beacon-based analytics

3 Likes

That’s like asking two people with wind-up clocks what time it is. Their responses should be close, then you can obsess over differences, or totally panic when something really doesn’t line up. Then you can get a third opinion that doesn’t agree with the first two, and so on.

I’d say just pick one that tracks what you want, accept it as a good enough sample size, then watch it like a hawk for trends and changes.

1 Like

Sorry if it wasn’t clear but when I wrote “If you want to compare beacon-based analytics” I meant consider using the Cloudflare WA, which will give results close - but never the same - as Google Analytics.

There is no way two different solutions will ever give the same number as conditions - network, privacy blockers, ad blockers all treat each differently.

The results will be close enough though and if you are interested in privacy for your visitor, better go with Cloudflare WA.

But never compare Google Analytics with Cloudflare domain analytics. Completely different beasts and the numbers will never be even close to being in the same magnitude.

2 Likes

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.