The content had a developer-y tinge (as you'd expect at Google's annual developers' conference) and a strong mobile focus.
The main topics discussed were:
- Stitching together cross-device visits via User ID
- Cross-Device Measurement
As you might have guessed, these features are all baked into Universal Analytics, so this is the first time at I/O they've been talking about the new product. You can watch the whole thing here.
Device Overlap With Cross-Device Measurement
What I thought was most interesting was Cross-Device Measurement, which lets us segment users based on multiple devices, in much the same way that Multi-Channel Funnels allow us to segment based on multiple sources.
Here's how Andrew Wales introduced the idea:
In this example, a customer visits on a tablet device and then later purchases on a desktop device.
Under the "old" Google Analytics, that would be reported as two Unique Visits: the first, on a tablet, has $0.00 revenue because no purchase occurred. The second, on a desktop, was a visit with a purchase, and has a $3.99 revenue attributed.
With Universal Analytics, it's possible to stitch the two visits together via User ID. This gives us a better understanding of how usage of different devices influences purchase and other user behaviour.
This level of tracking also leads to... this!
What this means is that we're soon going to be able to analyse Device Paths, just like we analyse Conversion Paths today.
Take a look at how the Device Paths report will look when it's ready:
It's just like Conversion Paths in Multi-Channel Funnels, except with devices. And, of course, I'm assuming that we'll also be able to segment by a combination of channels and devices for high-octane, 24-hour full power no shower data analysis!
Putting This Data To Good Use
In the talk, Andrew Wales homes in on how this data can be actionable:
In this example, we can see that users that visit first on Desktop and then on Tablet are very highly engaged, and spend a lot of money. Andrew says that this insight might make us decide to place something on the homepage as served to desktop users that encourages them to make a shift to their tablets.
Great for conversions, but also great for users -- we all want to use our tablets any chance we get, so if your site looks and works great on tablet, tell us!
Better Device Attribution
Finally, I thought this was neat -- Acquisition Source, giving extra conversion data for devices:
This is neat because instead of having analytics tell you that users on a smartphone don't purchase, it's telling you that users on a smartphone do purchase, but on other devices (or Touchpoints.) It's a little like multi-channel funnels telling us that actually, display advertising might not be the sucky cashsink that we all thought it was.
About Them, Not Us
It's difficult not to view this as being analogous to cross-channel measurement, and it's fair to say that there are similarities between the two. And certainly, this promises to be just as big a leap as multi-channel (particularly if we mesh the two sets together to create multi-device-multi-channel).
What I think makes this especially awesome is that multi-channel measurement is about us -- our advertising, our campaigns, our brand influence -- this is about them. It's about users, the way they choose to access content, and the way we need to adapt to their needs and their ways of doing things.