Discrepancy in Oculus DAU and Unity Analytics DAU

Unity Services
1

Hello,
I currently am working on a Gear VR app that integrates Unity Analytics.

As it is Gear VR, it has Oculus analytic data as well.

However, we are noticing a discrepancy in the DAU data between the two. We had thought it could perhaps be a delay in the data processing, but even accounting for a delay, the numbers do not add up.

A loose example:
Day 1: Oculus DAU 77, Unity DAU 36
Day 2: Oculus DAU 21, Unity DAU 44

Has anyone run into anything similar that could account for the DAU differences?

Thanks!

2

@InsideDown ,

Would you be able to submit a support ticket, and if possible could you include your Analytics project ID and a link to your game for testing:

3

Thanks ap-unity- I did submit a ticket and received an answer back; though at this point the answer is “we’re not sure”.

Could you tell me how a DAU is defined? Is there a time limit (must run app for X minutes) before a DAU is triggered, or is it immediately on app launch?

4

@InsideDown ,

Yeah, I saw that ticket and thought it might be a second data point that hints at an underlying issue with Oculus. But if not, then we can focus on your app. Would you be able to add a link to your project in that ticket?

Yeah, it should be at app start, if there is an internet connection available. It’s possible there might be a slight delay, but not several minutes.

If you use a network traffic analyzer like CharlesProxy, you can see the appStart event that is sent out.
https://support.unity3d.com/hc/en-us/articles/115002917683-Using-Charles-Proxy-with-Unity

5

Thanks @ap-unity , yeah, I’ve used Charles to copy that info and can see it firing. Will reach out to Oculus and report back with what I can find out.

1 Like
6

To those that may be searching for the same answer I was.

We seem to have identified the main issue. Oculus tracks based on PST (a daily active user is defined as starting at 12am to 11:59pm PST), while Unity tracks based on UTC (12am to 11:59pm UTC).

This would account for discrepancies between the numbers - it doesn’t fully answer everything, but it does basically say “these numbers will never be the same”. In general, we’ll just try and use both analytics to look for patterns/trends vs a direct comparison.