(I'm re-posting this from my own blog, BullishData.com)
At IBM, we talk about mobile technology a lot. Between the launch of MobileFirst two years ago, or the long-established Holiday Benchmark Series over the holiday shopping season, or the acquisition of mobile push messaging company Xtify just a few months ago, there’s a strong focus at IBM on developing and integrating mobile technology in the enterprise, and on helping businesses build better relationships with their customers via mobile interactions.
Anyone working in marketing or digital analytics can tell you that customers are increasingly using multiple devices in tandem today. Here are just a few examples from my own life:
- I browse on a clothing retailer’s website on my tablet before converting on my laptop
- I look up a restaurant’s menu on my work computer and check in there with my smartphone later that evening to secure a discount promotion
- I tweet about a brand on my phone, use their tablet app to find some information a little later, and finally convert on my laptop
These are increasingly common behavioral patterns we see, but they can prove to be incredibly frustrating for digital analysts to measure. The problem can be described by any manner of sewing analogies: “stitching” together each these interactions on different devices; “threading” them into a single customer’s experience; or simply associating them together into a single visitor profile, can be a tricky measurement problem.
The problem lies in visitor identification. How do I know that session 71936 (a shortened example of a visitor’s session ID recorded by your digital analytics tool), which came from a Samsung Galaxy 4 at 18:14 on May 18th, was the same person as session 91840, which came from an Apple MacBook Pro hours later that same evening? In the same time period, this site probably has hundreds of other visitors as well, with no immediate way to distinguish many of them.
The way we’re able to tie a certain session to a specific user is by getting him or her to identify themselves in some way. Note that I didn’t say “authenticate.” That actually has a more specific meaning that I’ll get into here in a moment. The bottom line, though, is: If you want to know who your users are, just ask!
On a mobile app, a popular way to do this is by prompting the user to provide login credentials when s/he first downloads the app. Your analytics tool will either save a visitor ID specific to that user’s app, or the unique identifier number for that particular device (both methods can work), to the user’s account on your site, allowing you to associate their behavior on the app to the visitor profile your digital analytics tool has already built.
On a mobile website, the same principle applies. Again, asking your visitor to log in (the same way they would on your website) is the easiest way to then associate their browsing behavior to their visitor profile in your analytics tool. One good way of doing this is to build an easy login option into your mobile-optimized site to retrieve a cart, or list of favorite items, or redemption code.
Without actual visitor authentication (i.e. your standard prompt for a username and password), you’ve still got some options. One of the best works with email. Via your marketing platform or ESP, you can basically include a identifying code that is tied to each individual customer when sending out messages (we call it a registration ID at IBM, but could be something different with another tool). That way, when I click through on it on my smartphone and I’m brought to a website, I’ve just identified myself, and your analytics tool can now associate the cookie ID on that device to my visitor profile!
What’s interesting, though, is that you only need to convince your user to authenticate one time on a particular device. From that point on, it is a straightforward matter to incorporate sessions from that mobile device with that specific visitor profile.
A cool feature that’s unique to IBM Digital Analytics is the ability to “reverse-map” previous visitor behavior after an identification event. That means that after recording multiple anonymous visits from a mobile device, once a visitor identifies her/himself on that device, Digital Analytics then automatically associates all of those previous mobile sessions to the same visitor profile, without any extra steps. In one fell swoop, you suddenly have a substantially richer history of that visitor’s total interaction with your site across all of their identified devices.
That’s some good-looking stitching, wouldn’t you say?
That’s really interesting, because associating and understanding behavior across devices and time is pretty much critical to understanding your visitor (not to mention a customer). I don’t know anyone who only uses a single device anymore – do you? As companies get better at building relationships with their customers across multiple devices and in different contexts, being able to measure the sum of those interactions will mean the difference between a meaningful message or an irrelevant one.
I, for one, get plenty of the latter already, thanks. The only way for marketing to now get my attention (and not a delete/unsubscribe click) is relevancy. Period.
I recommend that you also check out these other outstanding approaches to cross-device tracking by Adam Greco and the folks at Anders Analytics. Do you wrestle with this challenge too? Leave a comment below, and let us know what worked for you.