Why is Cross-Device Tracking So Darn Hard?


Share on LinkedIn

More people are accessing the web from multiple devices. This isn’t really news any more, as smartphones and tablets and the Internet of Things and whatever comes next keep improving in functionality and climbing in usage. The design world has been adapting in the last few years, with responsive design allowing site owners to deliver an optimized user experience to a number of device sizes and shapes from a single code-base. But it’s one thing to serve an optimized page to multiple devices; it’s quite another to tell whether those multiple devices belong to one person or to many.

Web analytics tools have had this problem for a long time: “unique users” has never truly lived up to its name. But with so many more people using so many more devices, and for those of us in the media world, with so many targeted advertising options available and at higher rates, distinguishing between unique people and unique devices has taken on new urgency.

Why is cross-device tracking important?

For e-commerce, it’s pretty simple. When you can tell that a shopper on her smartphone has made purchases with you in the past and even has left items in her cart earlier today, you can improve the user experience by allowing her to bypass certain functionality, like re-adding those items to cart or re-entering credit card information. Your product suggestions are likely to be better as well because you understand her buying habits, affinities and preferences. Personalize and simplify; convert and up-sell.

For media companies, it’s a bit murkier. All of your content is available to any user, whether on desktop, tablet, mobile, smart TV, or whatever else. You serve ads and collect revenue on each of these platforms. So how do you improve on something that it isn’t broken?

First, you need to know what the user wants (and doesn’t want) to see on your site; it will help you give her a better experience and keep her around so she consumes more content and sees more ads.

Second, not every ad is created equal. That’s basic Online Advertising 101: the more information you have about a user, the choosier you can be about what ad to show that user, and the higher the CPM you can charge for that choosy ad. You may have access to a mobile user’s previous mobile behavior, and to a desktop user’s previous desktop behavior. (Even that isn’t guaranteed, of course. There are folks out there like me who browse in Incognito Mode or Private Browsing by default, on both desktop and smartphone. I am part of the reason you’re overstating your traffic volume each and every month.) But if you can connect my behavior and preferences across visits, across browsers, and across devices, then your options for what you can serve me are far more lucrative than plain old display ads.

Why is it so darn hard?

Oh, no reason, just that virtually every means of identifying a “user” in the history of digital analytics has been based on the hardware and software (the device and browser) instead of the person. It couldn’t really have happened any other way, of course. A person doesn’t connect to your servers to request content; a device does. Tracking mechanisms have to work at that level, and they all have flaws when trying to translate into the world of people:

  • One person can map to many IP addresses and one IP address can map to many people, and IP addresses change all the time.
  • Cookies exist for a single browser on a single device. The cookies I send to your server when in Chrome are different from the ones I send from Firefox, even if I’m using both simultaneously.
  • Cookies don’t easily translate between a mobile app and a mobile browser, even on the same device.
  • Cookies expire. I can clear them manually any time I want in any major browser. I can set them to disappear automatically after N days. I can browse in Incognito Mode or Private Browsing Mode so they disappear after every single session.

The Internet just wasn’t built at the “person” level. That’s why cross-device tracking is so darn hard.

Technological solutions that probabilistically match different devices to each other are evolving. Companies like Tapad and Drawbridge are developing ways to link behavior on multiple devices to an individual user, potentially offering deeper and more relevant info about that user’s identity and preferences to you so you can more effectively target and personalize.

But even that won’t get you all the way there. Match rates will undoubtedly improve from where they are now, but exactly how good they will ultimately be remains a question. Even as efforts to map multiple devices to individual users improve, the device (not the person) remains the lowest common denominator.

What do we do about this?

The best way to identify users across multiple devices is, and likely will be for a very long time, to have your users log in.

Registration is, of course, the necessary first step. Your site has to offer some valuable reason for users to take the time to part with a little personal information. Why should I, the short-attention-spanned, ad-averse, distrusting user that I am, give you any info at all? What’s in it for me? Personalization, commenting, exclusive content, games, chat sessions, sweepstakes? Anything?

If you don’t offer any features behind a registration wall, I won’t register.

But even if you do, make sure that they’re available and appropriate across platforms. If I register on my laptop so I can comment on an article, but I never have the desire to comment from my Android because I don’t like angrily swiping my thumb across the screen that much, I won’t log in from multiple devices.

And, of course, if my cookies ever expire on any of my devices, I’ll have to log in again for you to re-associate my device with my preferences. If I only wanted to comment that one time, and there’s nothing else you offer me, I won’t log in again.

But if you can make me want to log in from different devices, by tailoring the experience so I see more of what I want and less of what I don’t, then you’ll be onto something. I’ll be more inclined to stay on your site longer, whether on a full desktop or on my smartphone, so I’ll see more ads. And because you know more about me from my behavior and from info I gave you during registration, each and every one of those ad impressions can bring you more revenue.

Randy Bird
Randy helps his clients improve their marketing initiatives by utilizing the wealth of consumer data they have at their fingertips. He has previously worked in multivariate testing and audience measurement with SiteSpect and with Compete.com. Randy holds an M.A. in cognitive psychology from the University of Illinois.


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here