It’s been just over a month since the General Data Protection Regulation came into force, and as far as I’ve seen, the impending doom that some predicted has yet to hit. Aside from Ghostery’s failed email, subject line “Happy GDPR Day – We’ve got you covered!” where they CC’d hundreds of user emails rather than BCC’ing, I’ve not heard of too many others. It’s OK though, because according to Reuters, the regulatory authorities were struggling to get ready in time as well.

What’s clear to me though, is that there are still a number of question marks about how and where brands should be collecting and using certain pieces of data. When we think about that from a data and analytics perspective, this is really important.

Understanding performance

For those of you who have joined one of our 2018 Future Focus sessions, you’ll remember the topic of the convergence of brand and performance marketing and how in 2018, more than ever, these two worlds are, in the words of David Ogilvy, “colliding”. Why is this important? Well, in today’s world of digital advertising, we have the ability to track and measure the entire consumer journey end to end, in more detail than ever before.

Not only does this give us insights into who consumers are and how we can best target them to drive a better brand experience, but through attribution, we have a greater opportunity than ever to understand the value each touchpoint gives on the path to conversion.

However, post-GDPR, two key players in the world of digital have thrown a curve ball that leaves us in an area of uncertainty around what comes next, at least in the short term.

Who are you talking about?

When you think of digital advertising, whether you’re inside the industry or not, two players that are most likely to come to mind first would be Facebook and Google. And as two of the top 10 most valuable companies in the world, and the majority of their earnings coming from advertising, their impact on the market is huge. And unfortunately, when they make a change, especially as significant as these, a bit of panic sets in.

Tracking Users in DoubleClick

First to Google. Globally, almost 60% of brands using ad servers use DoubleClick Campaign Manager (DCM). In the Australian market this number increases to almost two in three so when, just 30 days ahead of the GDPR live date, Google announced a change to its DCM log files, brands and agencies started to scramble.

The DCM log files (or data-transfer files) give an impression-by-impression, click-by-click and conversion-by-conversion view of a brand’s marketing activity. Rather than consolidating reporting into aggregated buckets, these files split out each action individually, giving information such as creative, publisher, timestamp (to the micro-second) and most importantly, UserID. When we are building an advanced attribution solution, the UserID is the most important factor.

The reason for this is that unlike econometrics, advanced attribution models build out each user journey independently of one another to get the view of the real experiences consumers had, be they converting journeys or not. Therefore, to build these, you need the UserID to ensure you’re confident that you’re only looking at a single user.

With GDPR going live on May 25th, Google stopped the tracking of the UserID (and PartnerID) for EEA consumers on that day. As part of their announcement, they also said that these changes would be rolled out globally, but as yet, there’s no update as to what that timeline might be. All we do know is that we need a solution in place before it comes this way.

Measuring impressions (at all) on Facebook

With Facebook, measurement across the entire user journey has always been more difficult than with other platforms. Aside from the fact that the majority of interactions on Facebook take place on mobile devices in-app, Facebook have historically not allowed the third-party tags (such as DoubleClick) to track impression data for any custom-audience activity. Given that one of the benefits to running activity on Facebook aside from reach is the ability to use the available customer data to target specific consumers, this means we miss out on a large part of the social picture.

Until recently, brands were able to track non-custom audience activity using third parties to get some additional detail into the customer journey across different platforms and publishers. However, that all changed around the 25th May as well. At this point, Facebook removed the ability to use third-party tags to track any impression data, taking away the ability to build out the entire user journey, end-to-end.

While we know that a user being served an ad doesn’t necessarily mean it had an impact on them and a click should have a much bigger impact on the likelihood to convert, impressions still provide value in understanding the consumer journey and when we build attribution models, it’s important we track consistently and accurately with as many touchpoints as possible.

Why now?

There are a few opinions on this one but I’ll give you two. The first is what was cited by both Google and Facebook in that they are putting customer data back in the hands of customers themselves. With both companies having the global scale that they do, there are concerns around the tracking of a UserID (in the case of Google) and use of third-party tags (for Facebook). In fact, for Google, GDPR defines such online identifiers as Personally Identifiable Information (PII), though only when the data is in its raw form and un-hashed or unencrypted.

The second thought though differs quite drastically. Rather than it being around the legalities, it’s a chance for walled gardens to get a greater advantage. For Google, while the UserID is no longer going to be included within log files, it will remain in their internal systems and, as outlined in the next section, can still be used, as long as it is within a Google platform. What’s interesting is that other ad servers, such as Sizmek, have come out publically saying they will not be removing the UserID from their log-level data, thus the GDPR argument seems to harder to believe at this point. What Google does have that Sizmek doesn’t though, is a much broader portfolio of products that utilise this information, so scale would be an issue in their decision making too.

For Facebook, with the Atlas experiment being unsuccessful, this move seems to be inherent of a power play to avoid the likes of Google being able to determine the success of activity run on its sites. Whilst Facebook do have their own attribution model built within Facebook Business, this is a siloed view of the world given it only tracks activity across Facebook platforms, giving more of a snapshot of the consumer journey than an end-to-end view.

What can be done about it?

While the Google changes will likely have an impact on the ability to accurately build a view of the consumer journey for attribution, there are a few options available. These options are:

  • Use a third-party tracker:

Using a third-party tag that tracks a separate UserID would mean that this identifier could be combined together with the log files to once again have a single view of each consumer. Some advanced attribution providers already tag activity using their own tags to reduce the requirement on the ad server to do so, which removes the issue. The biggest concern about this though is that Google recently changed the tracking capability on YouTube so only verified third-party tags would work there. This adds another level of complexity to this option.

  • Use Google’s Ads Data Hub

As mentioned above, Google will continue to track the UserID within its own platforms. Ads Data Hub allows you to look at the log-level data, and assuming you are able to, run the algorithms inside the platform. Despite the platform holding the data you need, you can’t remove information to use elsewhere unless it has been aggregated up from a log level. The stipulation for this option is that as a user, you must be certified in the product to use it.

  • Switch ad servers

This is the most drastic option of all, but given other ad servers have mentioned that they’ll continue to track UserIDs, it’s still an option. As long as you’re not too heavily invested in the Google product portfolio (i.e. through Analytics, Optimize, DBM, etc.), this could be something for your brand to investigate to see what is the right solution for you.

When it comes to Facebook, the opportunities are more limited. Unfortunately, Facebook is blocking the use of most third-party tags to capture impression data. However, for the purposes of attribution, Visual IQ, a global advanced attribution provider, has access to the log-level data with its tag and as it happens, have recently launched in Australia (just in time to help solve this issue if Facebook is a problem for you).

What about my brand?

With all of this, it’s important to understand that there’s no single solution that will be right for all. Each and every brand is in their own position in the marketplace with regards to data and technology, so each needs to be analysed as such. While the Facebook change may be impacting your brand today, Google are yet to announce when the UserID removal will occur outside of EEA consumers, so it seems as though we have at least some breathing room before this happens.

If you think these changes are going to have a significant impact on your performance tracking or use of attribution, please feel free to let us know. We’d be more than happy to talk you through the situation in more detail and make sure that we can find the right solution for you.

Shares
Tom White

Tom is the General Manager of Data Intelligence for iProspect, responsible for driving the strategic direction, delivery and performance of the agency's data offering as well as finding innovative uses of data solutions to give iProspect clients an edge in the marketplace.