The Confusion with ‘Cross Device’ Retargeting


The increasing number of devices users are accessing the web through has shaken up the retargeting game. The path to conversion is now split over a number of devices, with desktops, smartphones and tablets all potentially figuring in the path to a single conversion. Research from comScore shows that 67% of users start a purchase on one device and finish it on another.

When it comes to solutions to the problem, the term ‘cross device’ is often cited, yet different companies use the phrase in very different ways. Does it simply mean the ability to retarget on every device including mobile? Or does it mean finding the same user across different devices, and serving fully personalized ads on them all? As in all matters programmatic, the devil is in the detail. It’s time to make clear what cross device retargeting is, and what it is not.

Let’s look at the first option. Retargeting on every device is the ability to show retargeted ads solely on the device the user originally used to browse. For example, if a user browsed for shoes on her laptop, she will not be shown ads on her smartphone when she browses the web later in the day, or on her tablet in the evening. So, the user will see ads on her laptop but not across all of her devices.

The second option involves knowing which group of devices belong to one user, and showing them personalized ads across all of them. This time, after browsing for shoes on her laptop, our user is shown ads featuring shoe recommendations on her smartphone on the way to work, and on her tablet while relaxing at home.

Only the latter technology is true cross device retargeting. As Cassandra Bowman, Product Manager at Tapad, recently tweeted, “Cross device isn’t serving on different platforms, it’s finding your audience across different platforms.”

In our example above, the former technology only sees devices, and treats each one as a new user, ignorant of the fact that they may belong to the same person. True cross device retargeting sees users, and all of the devices that belong to them.

Why does this matter? Recall that the path to purchase is now split amongst multiple devices. This means that serious marketers need to use technologies that think in terms of users, not devices.

Retargeting that can’t identify users and the devices that belong to them has many issues associated with it. To begin with, it is impossible to properly cap ad frequency, as the retargeting provider will not know how many ads were shown to the same user on all of their devices. Browsing the same site on your desktop and smartphone will mean you get shown double the ads you should. This leads to an inefficient budget spend.

Secondly, the lack of proper frequency capping means the user will suffer a poor brand experience. They will see different ads on each device which, combined with the frequency capping problem outlined above, makes for a poorer relation to your brand.

Finally, the ads won’t be personalized to the extent that they could, or should, be. What if a user browsed for digital cameras on two devices, and ended up converting on the first device? They would still be shown retargeted ads for cameras on the second device even though they’ve already made a purchase. A true cross device solution would be aware of this, and show them ads for accessories, rather than showing ads for products that the user has already purchased.

The difference that true, user-aware cross device retargeting makes over non-user aware multi-device retargeting is significant. For a major US retail client, Struq have found increases of 19% in post-click conversion rates, and 72.1% uplifts in average ROAS upon the introduction of the new technology. Both periods featured consistent levels of spend.

It’s time to set the record straight on what cross device is, and what it is not. ‘Cross device’, isn’t simply retargeting on different devices – it’s having an understanding of the same user on multiple devices, and retargeting them accordingly.

Sam Barnett, CEO, Struq –

Copyright © Sam Barnett 2014; not to be used without permission




Please enter your comment!
Please enter your name here