We’ve got snowplow running on a site which has is using unbounce for landing page split testing.
The snowplow tracking code and the ga tracking code are both being run through unbounce. The problem we’re having is that there seems to be quite a large discrepancy between number of users & pageviews between GA and SP on some of the unbounce pages.
In some cases it’s up to 50% more, in others it’s up to 50% less. I’ve noticed that the GA integration for Unbounce seems to append the variant onto the page url, although unbounce support have said that the GA appears to be configured correctly.
Has anyone seen a problem, like this before with tracking on unbounce ?
So I’m not familiar with how unbounce works, but the instinct I have about the behaviour you report is that indeed something might be amiss with the tracking instrumentation. Specifically, I would say that it sounds like it might be to do with the timing of when the tracker script fires.
If it were related to a GTM instrumentation, I would probably be advising to look into the triggers for the tags, when they’re expected to fire, and what the priority of them are. I’m not sure if unbounce has these concepts, but perhaps that helps investigate.
In some cases it’s up to 50% more, in others it’s up to 50% less.
My gut tells me that some race condition might be at play on this - perhaps there’s a chance that pages are closed before both trackers have a chance to fire, or that the JS tracker sometimes isn’t loaded on time to register and fire the event?
That sounds like a good idea. That should (in theory) alleviate any sort of race conditions - you may miss some data from ad blockers that block GTM but that should be uniform across both Snowplow and Unbounce in that case.