No more cookies doesn’t mean no more cross-media audience measurement, but it does herald a new role for market research, suggests Henrik Lauritzen, CEO at AudienceProject.
Increasing fragmentation of media, walled gardens building their walls ever higher, Apple’s cross-tracking clampdown, Google replacing third-party cookies with FLoC, GDPR battles raging.
Understandably, all of this has led to concerns about how cross-media audience measurement will look in the future. How can you understand audience reach, frequency and composition across media and platforms with all these changes and regulations?
The task at hand is complex. There is an ever-increasing number of advertisement channels and access points; more walled gardens are springing up with increasing numbers of social (as well as other forms of) media platforms building walls around their inventory and, in some cases, audience reporting.
Robust methodology and technology required
Overall, it is a huge methodological and technical undertaking to build a measurement system capable of providing reliable measurements for reach and frequency in specific target groups across the plethora of media that are activated in a typical digital ad campaign.
But it is not impossible! For sure, the landscape places high demands on both methodology and technology and requires custom-built integrations with various platforms to be able to provide advertisers with cross-media audience measurement within the open web, as well as non-cookie ecosystems like Facebook, YouTube, CTV, and linear TV.
To temper the complexity, it is helpful that for the here and now, advertisers’ concerns seem to be primarily orientated towards understanding how to measure and improve effective branded reach across the fragmented TV/video ecosystem. This is driven mainly by the proliferation of (digital) video at scale and addressable TV, right at the moment in time when the (still powerful) linear TV option has entered a phase of declining reach to key audiences.
This video focus is helpful to get started, and will yield measurement solutions that will go on to be applicable to other areas of media investment, rather than potentially becoming stuck straight out of the gate by trying to solve all problems perfectly, at once.
And there is another perhaps surprising counterintuitive boon to the cross-media audience measurement quest in the form of Google’s decision to abandon third-party cookies.
A revival of market research
“How can this be?” we hear many people cry. Well, because it puts an end to cookie bean-counting as a means to measure digital campaigns, which (you will hear us cry!) has been hopelessly flawed for a long time. And at the same time, it paves the way for a revival of market research, which is all about dealing with incomplete datasets and inferring conclusions from these to the whole population – a practice that will yield far more trustworthy and actionable insight than just counting far-from-perfect cookies.
Not many people think about this, but so far, the demand for this capability has not been as big in digital as it was in more traditional media. The reason is simple. Why care about complicated statistical procedures when the general sentiment in the market often was ‘why bother?’ My ad server drops a cookie for every impression I buy, and that’s all I need to keep count of.
But how is one supposed to have the faintest idea about how many times and how many people in my target group are reached by a campaign running across multiple walled gardens, iOS inventory, and soon-to-be-cookieless Chrome inventory? Cookies might have got you part of the way, but it was with a reliance on a very imperfect and incomplete sum of something not intrinsically linked to any proven audience attributes.
So, our point is that cross-media audience measurement is definitely possible today, whereby the campaign is measured by a system built on advanced inferential statistics, graph technology and well-balanced panels and integrations. Enter market research for measurement in the post-cookie world.
Combining market research and machine learning will pave the way
Solving the measurement challenge today most certainly demands a new approach to technology and the inclusion of new methodologies such as machine learning in order to help with complex computations.
But at its core, it is still all about the good old statistics of classical market research such as weighting of panellists, extrapolation from census and sample to measure reach in the target group and more, only now embedded in and reinforced by powerful software. (As an aside, we believe it is important that machine learning is not presented as a ‘cure-all’ black box in this scenario, but rather that its utility is understood – so expect more on this in the future.)
The bottom line is that the death of third-party cookies won’t be the end of cross-media audience measurement. Think of it more as forcing a much better evolution of audience insight rather than continuing to misinform ourselves with an over-reliance on a flawed (but admittedly quite tasty) commodity.