skip to main content

Cookie-Based Test and Control

Dave EdwardsFebruary 20, 2014

The world of digital marketing has long been driven by the cookie as the universal standard and tool for brands to target consumers at the individual level. The problem, of course, is that a cookie is not an individual: one cookie could track the activity of multiple consumers and one consumer could have multiple cookies. This paradox drives a model that, while claiming to be data-driven, is predicated on data that may be incomplete, incorrect, and often corrupted.

Take the following as an example:

It has been determined that an advertiser wishes to advertise on Publisher Y. In order to measure the effectiveness of their campaign based on visitation rate and applications, this advertiser runs not only their standard campaign messaging but also a placebo ad in order to track both exposed and non-exposed consumers. This segmentation is set in the advertiser’s ad server with a 50/50 split, based on cookies.

While at work a consumer visits Publisher-Y, a website that they frequent quite heavily, their cookie-id is assigned to see the campaign message and they are exposed. This consumer is extremely busy at the moment so while they are interested in the offer, they cannot respond. Once home this same user once again visits Publisher Y however due to a change in device has a different cookie and this time is shown the placebo ad. This consumer now has the time to visit the advertiser’s site soon after this placebo view and fill in an application.

What will the analysis show? The analysis will show that one cookie ID was exposed to the campaign and did not transact, while the placebo group acquired a transaction—with no connection to a single user.

What does this mean to you? While cookies have long been the basic identification of the internet, its weaknesses are both well-known and pervasive. Cookies are not stable and unique to an individual. In a recent Forrester report, it was found that “…the third-party cookie, the most prevalent desktop-based targeting mechanism, is showing its limits at identifying and addressing your target audience — let alone at reaching the same user cross-platform.” When you also consider that the new editions of the big 5 browsers are designed for cookie opt-out as a default setting (and according to Shareaholic they own approximately 80% of browser usage), you now are not only working with a faulty approach, you’re now only basing your targeting strategy on a diminishing digital population, as people upgrade to these new browser versions. The Forrester report continued, “…Today’s targeting challenge for marketers is threefold: 1) reliably identifying, in a persistent manner, an individual consumer; 2) reaching that consumer across an ever-increasing number of devices and platforms; and 3) partnering with vendors whose infrastructure is still built around traditional targeting mechanisms.”

So what can a marketer do to address these challenges? We’ve found it is far more powerful – and insightful – to leverage technology that can truly enable personalized offers and ads to inform a test and control methodology to understand a campaign’s true ROI. Sound difficult? Too good to be true? In a few short weeks we’ll soon be publishing a point of view paper on our results doing just this type of analysis. In addition to our findings, the paper will offer recommendations to get the best ROI from your ad spend – because you’ll better understand the true effects of your advertising dollars.