Big Discrepancy between GA and GSC (>90%) No duplicate tags. Any ideas?

[Nutterum]

Hi Guys,

As the topic goes. A website I am auditing has a very strange issue. GA shows 90% more traffic than Google Search Console. Now, I know there are differences in the way the two systems count a visit as legitimate session, but I have not seen such big discrepancy between the two.

Do you have some ideas as to what might be the problem? Any tips you can share, on the times you faced similar problem? Anything that is outside the general, check for duplicate tags, make a Screaming Frog analysis, check for bots, etc. that you have used to solve or at least, reason why this thing happens?

Thanks for your answer in advance.

[Nutterum]

Just to update, I may have found the culprit. There is this strange archaic GA code on some of the pages that may be messing things up. I have fixed this and will update. Have you seen a GA code that has two separate scripts?

[goodroi]

Websites that have had many different SEO consultants can end up with a weird combination of code. To be safe don't forget to check if they had GA correctly installed (or uninstalled) on other domains. Some legacy sites can have old microsites that they forget to tell you about. I like to use my own scraper to make sure each url has the right code properly included and to also hunt down obsolete code that should have been removed.

[shaunm]

Have you seen a GA code that has two separate scripts?

Are you talking about the ga.js legacy library? I've seen sites using the legacy ga.js as well as the analytic.js together by mistake. I believe it's less likely that it could be the reason for the huge discrepancy in your case.

Btw, have you looked at your referral sources and set a filter to only include your hostname? I've read stories where people received fake traffic from sites that use your GA code on various other sites.

[FranticFish]

If the site uses subdomains and the legacy ga.js then self-referrals in GA could be inflating traffic. Unlike analytics.js, which does it by default, you need to configure the older code to identify this as internal rather than external traffic.

[Nutterum]

@FranticFish - Was about to update on that myself. Yes, a chunk of the traffic did came from such "scrap" traffic. Luckily we are almost ready with the new website, so hopefully with the universal GA code the traffic will be much more realistic (not to mention a pretty big redirect and cannonical list of URLs). Now its time to identify the exact amount of expected new traffic so that the client won't freak out on the sudden drop of numbers in his GA reports.

[Nutterum]

Update 2 . I have cleaned up the GA profile, however now the legacy GA code funnels organic traffic to direct traffic, thus messing up my reports heavily. I can still see the sessions coming from Google Organic as "source" however on the general acquisition tab 80% of the traffic is counted towards direct instead of organic. Working with Legacy GA code can be a headache...