Dec 11

Back to the future: Rebuilding trust in digital measurement

Advertisers regularly take business risks when investing in media.  Some choices need strategic commitment and yearly planning, and some also need dynamic and granular weekly optimization, especially digital media.  All need solid measurement data.  And in this era of slow economic growth, there is even more pressure to prove performance.

But there is a systematic measurement problem affecting digital planning across the industry.  This problem causes lower funnel or “performance” media to be overvalued and over used.  And upper funnel media like generic search to be neglected.  At worst it can lead to “cookie-bombing” consumers and putting them off.  Or it simply leads to bad targeting and wastage, hidden in low CPA rates.  This is partly because most digital attribution uses last click, which ignores the impact of most users’ engagement with media.  But even the more intelligent attribution techniques are flawed if they rely on cookies, for several reasons:

  • When consumers navigate using different devices, cookie-based attribution usually fails
  • Most interaction on Apple devices is left out because Apple defaults to disallow 3rd party cookies
  • Measurement is lost when consumers block or delete cookies
  • Impressions in “walled gardens” like Facebook cannot be tracked independently so are ignored
  • Powerful offline or external drivers of online conversion like TV campaigns or the weather are usually ignored

Media with persistent log-ins like Facebook and Google can get round some of these problems, but only within their own media.  And techniques that use predictive (probabilistic) modeling techniques to identify unique users also offer some help.  But the accuracy of these “black box” solutions will always be debatable, and the ethics of the data they use is questionable.

That’s why we adopt a measurement and attribution approach that does not rely only on cookies, and is designed to be available to all our clients live and on demand.  And it takes in to account offline and external drivers.  We use a combination of digital tracking, tried and tested modern econometric thinking, and cutting edge machine learning techniques.  We also provide insight into optimal spend and frequency capping at placement level.  And show where each paid, earned, and owned media is encountered in the user journey.  Finally we provide overlap analysis showing which media are targeting the same users.  All so that media can be optimized to deliver more conversions at less cost.