On my app, I noticed that I have a significant share of the visits that have no "page view" hit.
On the other hand, these visits have a "custom link" hit with "custom link = ADBINTERNAL: Lifecycle".
So I think that these visits are counted when someone open the app and then close it immediately, before to have a trackState event counting a page view (hence the Lifecycle but no page view).
But the share among total of visits seems high ... Anyone to confirm I'm right (or not)?
Also, I wonder what is the good practice in this case ... Should we count these visits without pageview in the global traffic or not? Because potentially, this has a significant influence on the conversion rate figures that I communicate!
Sounds like you're on the right path to figuring this out. Is the data any different when using Launches as your metric instead of Visits? My guess is that it wouldn't, as I believe it's a Lifecycle metric, but it's worth checking.
Good question about best practice here. I *do* think it's worth reporting on this issue (that a not insignificant amount of visits are canceled before anything even loads.
However, it may make sense to create a calculated metric for Visits with a visit-based segment applied that says "Pageview exists":
And use that as your denominator.
Also, you may want to continue digging into the other lifecycle metrics to see what is actually happening in these instances by creating a visit-based segment where Page Views does not exist, and analyzing it. Perhaps, for example, the app is crashing, and that's the root cause. It may also be interesting to see what mobile devices/OS version is aligned with this issue.
Well, indeed, it seems the "Launches" metrics is directly linked to the Lifecycle hit...
Your example of segment is exactly what I did to check:
- one segment "Visits with pageview" set as a "visit" container with pageview metrics "exists".
- one segment "Visits without pageview" rthat"s exaclty the same one but in exclusion.
I do also have some errors counting in custom links (so that can provide visit no page view), but very few compared to the total volume of visits with no pageview.
So I guess the explanation of "people opening the app (maybe by mistake ; or clicking on a link opening the app...) and closing it right away before to have any real page view" is still the good one?
I'm just amazed by the number of time it happens! I have no idea if it's above or under average for an app...
And to conclude, what I understand is that there's no "best practice". On one hand it's interesting to have this figure (to monitor it and make sure it's not increasing too much), but on the other hand it's also interesting to exclude those cases to have a "useful" traffic on which calculate a proper conversion rate.