Good to see Google Analytics App+Web move beyond GA for Firebase and toward delivering the promise of Next Generation Analytics. Enter session-level data!
Remember how only a while back we had to count instances of the
session_start event in order to get the count of Sessions? GA App+Web originally inherited the app approach of GA4F, where source of app install seems to be the most respected marketing piece of information, and users count more than sessions. This is slowly changing, as App+Web data comes a step closer to Universal Analytics with some session-scope data.
(While this is only arguably a change toward clearer understanding of user engagement with the website or app, it definitely is better understood by marketers. After all – we pay for clicks or sessions delivered….)
Sessions – this is pretty much equivalent to Universal Analytics. A Session is started with a
session_start and lasts until the user closes the browser (web) or the app (app). The timeout is 30 mins. In the case apps, it starts counting when the app moves to the background.
Engaged sessions – this is more or less opposite of Universal Analytics Bounces. An Engaged session is one with at least 2 pageviews, or a pageview + an interaction, or 1 pageview with 10-second duration. The 10-second timeout cannot be changed, for now.
Conversions – these are not immediately useful if we mean to duplicate Universal Analytics reports in App+Web: Conversions are a count of events marked as Conversions, not deduplicated on session level (so it is not really a session-scope metric). What is more, right now there is no way to specify which conversion we want to view – they are all thrown into the same pot. The most likely use case is when we only track a single Conversion event – then the report will be more meaningful.
Last click source, medium, campaign – these are a welcome addition to App+Web attribution data. They correspond to UA Source, Medium and Campaign. Voilà!
Last click Google Ads campaign (& other Ads dimensions) – more of these are available for Ads. And understandably so, it is Ads that pay for the party (mostly).
We may still feel some things are missing when moving from Universal Analytics – especially in the session-scope area. And so:
- Conversion breakdown – a separate metric for each Conversion would be more meaningful.
- Conversion deduplication – the whole beauty of UA Goals is that 2 form submits during 1 Session still count as 1 Goal Completion.
- Conversion rates – I don’t consider this a particularly meaningful business metric, but I can understand how it can be useful in quick assessment of effectiveness of campaign traffic.
- Landing pages – the dimension isn’t available right now. We can arrive at this data in Path Analysis, though. It is not very convenient as we have to use filters to get Last click source/medium etc. (no session-scope dimensions in segments just yet…) but the data is there:
- Engagement metrics – apart from Engaged sessions, we don’t have these: entrances, exits, timings.
How to work with session data in App+Web
When there was talk about App+Web changing the paradigm, it was more true than we might have thought. Are we ready to start looking at users’ interactions with our websites (and apps) as an ongoing experience, unsullied by the limits of Sessions?
Even if we are, our colleagues in sales and marketing are probably not. BigQuery may help here.
We can use BigQuery to get Conversion breakdowns and rates. But regarding attribution dimensions, right now BigQuery export does not include session-scope data: it is still the same as in GA for Firebase. Apparently, however, Google are looking at introducing session-scope data to BigQuery as well.
To sum it up
Obtaining session-level insights from App+Web is a beta moment for an analyst. We still need to wait before App+Web becomes a platform that meets the data needs of marketing. It is taking steps in the right direction, though, so let’s hold on as this is not the final stop yet.