Dear Technical and Internet Gurus:
Is Google Analytics accurate?
@Teri_Kanefield I find it’s become harder to use
@Teri_Kanefield accurate-ish.
https://checkmyads.org/ has done a lot of good work investigating Google's role in funding hatesites and misleading it's paying advertizers.
@Teri_Kanefield it's good enough to look at trends but not enough to tie its metrics to a business contract. Too many things can lead to inaccuracies.
@grumpasaurus
Sounds like a "Take it with a grain of salt the size of a cow lick" situation.
@Teri_Kanefield
@Kencf618033 @Teri_Kanefield ga4 simply doesn't make any guarantees in accuracy and most client side implementations of it are subject to client framework, firebase, mobile os, and browser plugin shenanigans.
To make overall marketing decisions or understand your acquisition channels and calculate CAC, yeah it's the best free tool out there.
If you're going to do something like an affiliate payout program, it's not the tool to use.
All I'm doing actually is getting an idea how many people are there so that when things go wrong on the website I can rule out an overload of users.
@Teri_Kanefield @Kencf618033 because ga4 only provides real time data within the past 30 minutes it may not be the right tool for a Network Operations Center type of situation. GA4 is a reflection of your client application activity vs say a malicious attacker trying to hit your endpoints directly.
It may prove to be a quick sanity check if you don't have anything else though. (Servers reporting high traffic but ga4 reports low users).
In general it sounds like google analytics under reports
@Teri_Kanefield @Kencf618033 out depends on what you're looking at. In general "full data availability" isn't available for at least 8 to 12 hrs. If that's not what you're referring to it can be the result of a multitude of things from ad blockers to overall how your client app was built
@Teri_Kanefield Forget it, Teri. It’s Googletown.
@jlundell what does that mean?
@Teri_Kanefield @jlundell it’s a chinatown ref
@Teri_Kanefield depends on what you mean by that. It can undercount a bit because it depends on a JavaScript which some people block. Also people can reject cookies. But, assuming you’ve set it up correctly, it’s quite good. One thing to keep in mind is the complete data doesn’t show up until the next day. If you’re going beyond just things like pageviews, user sessions, and referrals, there’s a little bit of an “it depends” if you get either really high or really low traffic. Like if you are looking at physical location data for visitors and have low traffic, Google Analytics may not show it to you bcs it would be too easy to track info about an individual. You’d see a little caution triangle at the top of your report and when you hover it would say thresholding was applied.
But this is getting in the weeds. I’d say it’s generally quite accurate and very consistent so if you compare traffic from one time period to another to understand trends, you can trust it.
@Teri_Kanefield I’m happy to answer more specific questions. I don’t know much about how it interacts with advertising, but do know a bit about the use case for blogs.
@Teri_Kanefield One more thing, if you're comparing your GA numbers and site logs they might be very, very different especially if you're posting on Mastodon (as you are). My Netflify site logs show orders of magnitude more more pageviews than GA, because Netlify is counting every single hit, including, it appears, every time the site preview is loaded in Mastodon. GA tries to only capture actual people visiting your site. It can still be spoofed by bots impersonating people, but except during some kind of botting outbreak, I've found that to be uncommon. Your milage may vary with a more popular site.
@dys_morphia @Teri_Kanefield I used to run a very heavily trafficked site, and GA doesn't match site logs even if your site logs are filtered for non-bot pulls of main html pages. That's to be expected, because the GA javascript has to load after the actual page loads, and there are myriad reasons (as @dys_morphia has mentioned) that the main html can load and never load the GA elements or let Google know that your page was loaded.
@dys_morphia @Teri_Kanefield People still block JavaScript? No way - there are hardly any websites that don’t use #JavaScript. Teri - the problem with asking techies a question, they get all serious and anal about it. My answer - Nothing wrong with google analytics. Please use it (from another Techie turned entrepreneur who has learned to fudge things ;) #GoogleAnalytics
@Teri_Kanefield All data can be manipulated. It provides good trend information but I wouldn't trust it with something like a political prediction.
@Teri_Kanefield Bots; so no.
@Teri_Kanefield No, not necessarily, because more and more people are using adblockers and similar to block tracking what they see and do on the web, and good ones block some or all of what Google Analytics attempts to track. When you combine adblocking with VPNs, google analytics may record no visit, or a visit from a different location.
@Teri_Kanefield Really depends what you're trying measure, and how careful you are drawing conclusions. Generally fine for getting a relative sense of traffic trends over time, or which pages get the most views, or what countries most of your audience is from, but even that kind of stuff can be wildly off in some cases
@Teri_Kanefield There are many of us who block G-analytics because:
As a web browser user:
- It is spying on us
- It really slows down web browsing
As an operator of websites:
- It is merely one of many tracking systems. It tends to be more synoptic as it uses various attributes (e.g. browser fingerprints) to link activity on our websites with activity on other websites (but we website operators don't see as much of that data as Google does.)
I am not sure whether the new generation of browsers blocking third party cookies will have any real, positive impact.
@Teri_Kanefield It is precise but is fuzzy in accuracy. Always use it for trends only and comparing against similar google analytics data sets. It shouldn't be compared against analytics gathered from another tool, for example.
Bots, adblocking, virtual private networks, there's a lot of things muddying the water, but it can provide some insights too, just not to precision.
@Teri_Kanefield
It can't be completely accurate. I block the Google Analytics javascript in my browser, and I can't be the only one.
@Teri_Kanefield Good answers so far. In short, it can’t catch all traffic and it’s getting less accurate with time as more folks use blockers and privacy-minded browsers. It’s also WAY too complex and difficult to use for most cases.
Use WordPress’s Jetpack or a more responsible service like @plausible (which I use on most of my sites).
@Teri_Kanefield How much accuracy is worth overriding people's privacy (those not already blocking it)? I use goatcounter.
@Teri_Kanefield Generally, yes. Very.
@Teri_Kanefield It’s accurate for the set of your users using Chrome. Anyone using anything privacy focused will block it. I used to use it on my own site, and it would never show my own visits even from Safari (I guess because of Apple’s toggle to block cross site tracking). I’ve ripped that google code out of my site now, because: Google.
@Teri_Kanefield It is if you're Google.