Hi there!

I noticed this as well. In our Community (www.a1community.net) the differences between GA and inSided Analytics are quite high - on average, unique users and page views are 30 - 50% higher at inSided Analytics compared to GA numbers.

I think that is has something to do with the cookie consent. Since we introduced it to our community, the numbers went like this.

But I’d be glad if someone from inSided shed some light on this!

Best,

Wolfgang

If I could get a penny for every time I compared these metrics…

Of course neither GA nor we do own the true value here - while Google surely is undercounting (mostly due to add-ons but also cookie consent, and lateley apparently also the Safari browser), we are slightly overcounting (e.g. we have a harder time to count non-logged in visits on mobile / desktop as a single visitor). I might be wrong with this but I feel that the general trend is that GA is being blocked by more and more users these days.

In general, I do see the delta here to be around 30% - 40% when it comes to visitors. Page views somehow can differ, for reasons not clear to me (yet). The way that you handle your cookie consent surely plays a bigger role here. I suspect that the difference is even bigger (up to 60%) if your cookie modal by default does not include GA tracking. Next to that, probably the target audience plays a role as well (e.g. young and techy B2B customer vs. slightly older B2C end user).

I hope you understand that for us it is also a bit hard to give general information about the difference, as this seems to vary a lot depending on the context of each community. We will keep an eye on this in our benchmarking as well.

Understand @Julian . What about unique registered users? Can we trust that one is accurate and how is it calculated? 

Thanks. 

Hmm I have not much experience with this too, as this is a relatively new KPI.

It probably simply counts the unique user logins (and their page views) for the selected period. I don't think these numbers carry much risk to be incorrect. I calculated the unique user logins myself quite often in the past - which in itself is an indicator that it is simple and very easy to do.   This probably is measured directly within our database, based on e.g. counting requests to our server.

A potentially more interesting question is: Do we actually measure Community Managers, Moderators and Administrators here too? Or is it limited to the primary user role Registered User? I’d assume that we at least also count users that are a Super User as well...

I will check on our questions with the team and get back to you with more details.

I’ve found success getting accurate counts by using the Salesforce integration. If you’re using it (and have it setup where community member records are near a 100% match rate with Salesforce contact), it tracks Community Visits*. 

We use that as our active user count and look at it daily, monthly, and quarterly. Much more on our approach in my reply to this question: 

That active user count is usually a 99% match to what GA gives us as Active Users. Although as a private community (for now), our visitor metrics are setup to only see logged-in users who access the community, so it’s easier to be accurate. 

*we use visits not login, because the login cookie often persists across multiple days, and we want to see daily active user trends. Since they have to be logged-in for that visit to be collected, visit gives us a better way to see someone who is active across multiple days with one active login event. 

a very relatable plight @Gabolino. I’ve always just kept a doc measuring both GA and inSided metrics, and tend to air on the side of caution by leaning towards GA metrics. I remember in the past that any user who landed on a page and refreshed it always counted as another pageview, but not sure if that has changed since. The SFDC data @DannyPancratz mentions, though, is a great call if thats in your tech stack. 

If unique registered users are calculated how many users have logged in within timeframe this number should be 100% accurate.

Is it?