We have seen a lot of change in the identity landscape over the past 10 years. What once was a bit of the Wild West has turned into a highly regulated space. Yet it remains fraught with bad data due to a wide variety of ad fraud, as well as limitations in tracking due to third-party cookie deprecation, Apple’s ATT and Android’s Do Not Track.
Even with the current limitations in tracking, however, companies tout billions of ID linkages in the US alone. One company says it has data on over 200 million households in the US, which is baffling, since, according to the US census bureau, there are only 131 million US households.
The industry is too focused on quantity over quality. Unfortunately, it seems there are many people still chasing scale due to a lack of understanding.
The need for comprehensive analysis
If you are still doing “lift” tests to see how much more scale you can get, then you are only hurting the performance of your platform.
The truth is this: the “scale” you are buying consists of a very large volume of IDs that are unactionable due to all kinds of ad fraud from adware, bots and click farms and temporary mobile IDs from iOS and Android’s Do Not Track. Building audiences and/or targeting ads to these IDs will only result in poor performance.
There is no doubt we have all seen the results of this in the downsizing of ad campaigns across all major platforms and even the bankruptcies of companies like MediaMath. Still, the problems persist.
How do you resolve them? Stop using lift tests to make identity vendor decisions and invest in a full-scale platform test.
In my experience, testing the identity capabilities at the platform level to measure the quality of the solution being provided has proven most successful. Then, building audiences off of those IDs and running media in true A/B tests to understand the quality of the identity data you are testing enables you to discover how such data will affect your platform’s performance.
Don’t take the easy route
Simply adding more IDs is not the answer. Adding more addressable/actionable IDs that are properly linked to the corresponding households or consumers is the only true solution. Testing and proving this is not an easy process and must be done at the platform level at full scale with a comprehensive performance analysis.
Not enough people are talking about quality when it comes to identity because it is such a difficult problem to solve. It requires a combination of deterministic data and AI to not only identify the IDs that are actionable but attribute them to the same consumer in a privacy compliant manner with a level of accuracy high enough to drive up performance.
It also requires a multitude of disciplines – from technical to advertising, marketing and finance – working together to measure the results. It is difficult to coordinate that level of cooperation across disciplines in today’s fast-paced environment.
Yet such an effort is the only way to prevent your organization from making a common mistake. Don’t focus on quantity over quality.
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow BDEX and AdExchanger on LinkedIn.
