Crawl, walk, swim – don’t get bitten by the cross device shark


Get the basics down before  attempting a cross-device strategy

Jos Pamboris

Jos Pamboris, Chief Product Officer at Flashtalking.

It wouldn’t be summer in the US without their TV Shark Week, but for the global ad tech industry it’s seemingly always Shark Week. That’s because, like a shark, the industry is constantly moving forward. Right now, the chum in the water is all about cross-device strategies. This is understandable, because consumers increasingly live a cross-device existence. But beware before you jump into cross-device waters with both feet. It’s best to have mastered a single device and all of its complexities before shifting your focus to cross-device. Otherwise, as Bill Gates once noted, “automation applied to an inefficient operation will magnify the inefficiency.”

The lure of Cross-Device

A cross-device strategy is tempting, but heed Gates’ warning. It may distract you from more challenging aspects of our industry. Those include viewability, fraud, a dearth of sophisticated fractional attribution methodology and ultimately a lack of a clear marketing message that creates true influence. Ads that are not seen have no influence. But most measurement platforms are still attributing the sale to these ads.

It gets worse. Many hucksters in this industry try to hoodwink marketers by playing fast and loose with cost-per-action metrics.  They’re telling marketers, “I can’t help you with eliminating rewarding conversions to unseen impressions which doubles your CPA, but I can sell you cross-device attribution which gives you a significantly decreased CPA.” In this sense, and in many cases in ad tech, the CPA is subjective. For example, I can also decrease your CPA by simply changing the post-impression window from three to five days. This shell game hasn’t created any more real conversions.

To make matters worse, CPA is often hinged to the last click/impression attribution model. That’s a deeply flawed method. It’s the equivalent of the bottom-feeder claiming 100% reward for the kill. The truth is that it’s a collaborative effort.

Question the Device Graph

There are two methods to create a device graph: probabilistic and deterministic. While not mutually exclusive, an ad tech provider will generally favour one method over the other. Probabilistic methods take into account multitudes of data points and link them to specific algorithms to determine the probability that two or more devices are owned by the same user. Geo-location, IP addresses and many other data points make up this probability.

Deterministic methods depend on signed-in accounts (social media platforms, emails, phone numbers, addresses etc.) hashed for privacy and linked to multiple devices.

Demand Data Ownership and Portability

Accurate and objective data is key. Cross device should be part of a marketing strategy that ensures each device you target is being properly measured by a trusted third party. A lot of ad servers are not addressing the fact that they have gaps in their own capability to connect a user to a single device. Take mobile for example.  In mobile, apps and mobile web are totally separate environments. Many cross-device offerings cannot connect the two. Both probabilistic and deterministic methods have their issues.

Probabilistic is far from 100% accurate and deterministic generally requires you to play within a walled garden. As the name suggest, walled gardens prevent the marketer from porting their campaign data elsewhere, including their own attribution or DMP platforms. These sealed communities will prevent marketers and publishers from better understanding how best to optimise  the campaign based on the cross-device attribution results. This data accessibility is also a problem when dealing with DSP or specifically sell-side properties that have a cross-device offering. It is very difficult to validate the accuracy of the cross-device targeting on offer.

As a marketer, you should be asking your cross-device partner to provide its device graph data mapped to campaign data back in as granular a form as possible (log files, for instance). Without this data portability, you can’t connect your work in cross-device from one provider with all of the other integrated ad tech platforms at your disposal.

Get the Basics Down

Ultimately what should cross-device measurement provide? It should provide an unbiased, objective view of the performance of your media activity. It should provide trusted data that allows for actionable analytics.

Advertisers shouldn’t tolerate any less. Luckily, advertisers are not like sharks mindlessly swimming forward to an effort to stay alive. You have the luxury to wade into uncharted waters more thoughtfully. In doing so, make sure you’ve mastered the fundamentals first. Otherwise, you might end up getting bitten on the rear-end.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s