Why would the GA data NOT match Moz reports?

[timesharecmo]

After carefully double checking dates, and views, I have a site that reports via GA an organic traffic number 2x what is seen in Moz. On top of that, Moz shows a steady decline in organic traffic (from all sources) yet GA data is rising.

As you can imagine when a client sees this they get nervous.  Mozzers tried to help but told me essentially there's nothing they can do and I should try disconnecting and reconnecting the GA connection.  I am not sure what will happen when I do this, but when I did it anyway, nothing seemed to change, so I have no idea what to expect from these organic search reports in Moz.

Has anyone experienced this? I have to say I'm now wondering if  GA is the issue, but I am stumped as to how to go about figuring it out.  We have audited the GA code thoroughly.  I suspect there's now some kind of issue with duplicate traffic on their site that might inflate sessions (eg does a 302 read in GA as one session, or two?) but then again, why wouldn't Moz report that data, even if it is wrong?

Has anyone else ever seen this phenomenon?  Google searches revealed nothing....

[Gyorgy.B]

Would you share your website? It's difficult to check for bugs without seeing the tracking code.

Also try this:

1. Disconnect your Moz account from your Google account.
2. Create a new Google account and give read permission to the website's GA property's reporting view
3. Connect this new Google account to Moz and wait for the results. Unfortunately Moz won't pull in the historical reports, so you'll have to wait 2-3 weeks to see the results.

[timesharecmo]

Thanks Gyorgy,

As this client is a large public company I can't really post their code but I can tell you it's installed properly.  Anyway, even if it wasn't, if Moz is pulling this data from the GA API, shouldn't it match perfectly?

Melinda

[timesharecmo]

I may have found my own lead on this which appears to be some limitations in Google's API.  this is a very, very large site to do a URL parameter proliferation issue creating millions of pages on the site. I bet it's safe to assume that Moz has to use sampled data past a reasonable point, and given the size of this site, it's fair to consider the possibility that Moz couldn't match the internal data as a result.

Will report back as we work with them on reducing duplicate content and see if that pushes things closer together.