Is retargeting really effective

Retargeting despite (t) data loss

Awareness and traffic campaigns in display advertising are often refined through appropriate retargeting measures for the users already addressed. But what to do when the data pool of “collected users” is getting smaller and smaller? Cookie banners, a lack of user consent and developments at the browser level are currently making it difficult to track users and causing data to disappear. But part of this lost data can be used effectively to control and optimize retargeting campaigns: the lack of consent.

For some time now, data-driven forms of advertising have been struggling with a dwindling database, especially cookie-based technologies. The creation of user pools for retargeting in particular, but also the measurement of success in general, are becoming more and more difficult.

The restrictions surrounding the GDPR and the ECJ's decision on cookies ensure that less and less data is available for retargeting and for performance control. In addition, issues such as browser cookie blocking are of course also playing an increasingly important role. Whether Safari (ITP), Firefox (ETP) or in the future also Chrome - sooner or later all browsers will be blocked by standard third-party cookies.

The catch with all future technologies

There are undoubtedly numerous approaches how to meet this challenge in the future, because the days of cookies and thus all alternative user identification methods such as fingerprinting & Co. are numbered. Server-side tracking etc. are hotly traded solutions, however, and this is the current challenge, they are not (yet) available in a mass-compatible manner.

Cookies are currently the standard and we should handle them as best we can. Drawing the optimal conclusions from the current data situation is the real challenge in everyday life, despite all future orientation.

The blind spot in the data image

All of these factors make us blind at first glance to one eye of the data. We can simply no longer see some of the users on our site, track them and accordingly cannot address them again via retargeting.

But it is precisely this blind spot that we can take advantage of.

Step # 1: Form user clusters

With all the negative effects that the implementation of a cookie banner has on our data pool, there is one very decisive factor that we can take advantage of here: the clearly definable point in time from which the banner affects the Tracking has - the time of installation.

From this point on, our data image is divided into two clusters:

  • Cluster # 1 contains all users who have given their consent to tracking via the cookie banner.
  • Cluster # 2 includes all users who do not want to be tracked, i.e. who have not consented to the cookie banner.

In order to be able to use both clusters to address users again, we are creating user pools at two points and combining them with one another:

  • User pool # 1: all users who have clicked on the ad (can be displayed directly in the ad server, user consent is also a prerequisite here)
  • User pool # 2: all users who have consented to tracking. The easiest way to realize this pool is to set a soft conversion with the consent of the cookie banner.

Combined, the whole thing now results in retargeting lists that can be optimally used for re-approach:

  • RTA list # 1: Users who could be tracked on the landing page because they gave their consent (user pool # 2 with tracking consent)
  • RTA list # 2: Users who clicked on a banner but could not be tracked on the landing page because they did not agree to the cookie banner on the landing page (user pool # 1 with a click on ad - user pool # 2 with tracking -Consent)

Step # 2: User cluster performance check

In order to draw conclusions about the basic performance of both clusters, we go one step further. We use the above-mentioned advantage of the cookie banner implementation, i.e. the knowledge of the installation time.

The performance of the clusters before implementation can be easily compared with the comparison period from implementation. If we now make this cluster comparison at different positions on the website journey, interesting conclusions can be drawn about the clusters.

Two simple examples to illustrate this:

Let us assume that the measurable traffic on the landing page drops by 30 percent in both examples with the implementation of the cookie banner.

  • Cluster # 1 is responsible for 70 percent of the traffic.
  • Cluster # 2 accounts for 30 percent of the traffic.

In both cases, we also analyze the comparison periods in terms of conversions.

For example 1, the decrease in the measured conversions after the implementation of the cookie banner is only 10 percent. Cluster # 1 is responsible for 90 percent of the conversions here. Due to the high proportion of conversions, it is clear that the still trackable cluster has above-average performance.

Example 2 behaves completely differently, here the conversions go down by 80 percent. Cluster # 1 is therefore responsible for a comparatively small proportion of 20 percent of the conversions.

These results allow specific conclusions to be drawn about the performance of both clusters. While in example 1 the performance focus is clearly on cluster # 1, addressing the users in cluster # 2 again could even be neglected in case of doubt. In relation to example 2, it would be a clear mistake to neglect cluster # 2, since a relevant part of the performance comes from this cluster.

In addition to comparing the clusters of traffic and conversion levels, every single step in the website journey can of course be evaluated in exactly this way. Just as we have set the cookie consent as a soft conversion and combined it with the click on the ad, every other interaction of the user on the website can also be defined as a soft conversion and evaluated according to the same pattern. There are almost no limits to the level of detail.

All of these user pools can be used perfectly as a retargeting target group!

In the future, there will probably be completely different solutions for addressing user groups, optimizing campaigns according to objectives and all without the much-criticized loss of data.

However, this mass-market solution does not yet exist. Pragmatic approaches help not only to admire the problem, but also to change perspectives and derive conclusions for optimization from the loss of data. This approach is one of them - because data loss should not deprive the desire for display advertising!

About the author:

Nico Loges has been active in online marketing since the late 1990s. Initially, the marketing of musicians was the focus of his work. At the beginning of 2011 he dedicated himself completely to performance-driven online marketing and was responsible in-house as “Head of” all marketing channels of one of the largest German online retailers for fasteners and hardware. The move to the agency side followed in mid-2014. At web-netz, Nico Loges looked after customers in the areas of paid search and display. As head of the Paid Media Marketing department, he has been responsible for more than 25 employees since 2019. As an author, he is responsible for publications in the textbook "Guide to Data Driven Marketing" and in the agency blog. Furthermore, Nico regularly gives lectures at conferences and shares his knowledge in web-netz webinars as well as external workshops and other training formats.

EVENT TIP ADZINE Live - Advertising IDs - State of Data Nation? on June 9, 2021, 3:00 p.m. - 4:30 p.m.

The days of addressable web traffic based on third-party cookies are numbered. Due to restrictive browser settings and the consistent implementation of the requirements of the GDPR, alternative concepts are needed so that advertising continues to find its way to the right target group. Most of those involved on the supply and demand side in the digital advertising industry have great hopes for (universal) advertising IDs. Experts from advertising marketing, agencies and technology provide us with answers to the countless questions in connection with the use of advertising IDs and explain the current state of affairs with regard to the practical use of advertising IDs for media buying and marketing on the publisher side. Register now!