With the end of cookies and Mobile Advertising ID’s fast approaching, there is no shortage of discussion about ways to maintain some form of user identification. We need to figure out how data-driven advertising will work within the scope of privacy regulations and technical possibilities. Lately, this debate has revolved around two very different strategic approaches: shared user IDs and cohorts.
Shared user IDs
Shared user IDs aim at solving three main issues related to third-party IDs (and their disappearance):
- The inefficiencies related to cookie matching and the drop off between platforms using different identification framework. This renders up to 50% of web-based user data and inventory to be unusable.
- The need for user IDs to be accessible to 3rd parties to the main web server. Instead, these will be replaced by a shared first-party ID, which the server itself provides.
- The need for greater control around the usage of IDs and associated data, particularly concerning privacy preferences expressed by users.
On paper, shared user IDs are a very efficient and privacy compliant replacement for cookie-based IDs. Their main challenge is adoption by publishers and technology platforms as a scalable alternative to third-party identifiers. Like any currency, their value grows exponentially with usage. This implies a limited value to start with and a difficult take-off (the infamous “chicken and egg” challenge). The next few months will be critical for shared identification solutions to establish themselves as viable and scalable alternatives to third-party identifiers.
On the other hand “cohorts” approaches aim at grouping users into segments to avoid user-level data distribution. They have become a topic of discussion following Google Chrome’s Privacy Sandbox experiments. Cohorts have even been presented as the solution to privacy concerns. They are not.
Cohorts are merely a way to avoid the problem. The approach aggregates people’s data to pretend that we are treating them as groups rather than individuals. However, no regulation prevents marketers and publishers from addressing individuals. Instead, it requires those individuals to be informed about the data that is collected on them, how it is processed, the purpose of that processing, and to say whether they are ok with it.
A personal approach
Is it better to harass a group of people who have seen a particular product on a website until they buy it? Or is it better to engage in conversation with a single person to understand whether they are truly interested in the product and should be shown the ad?
Personal data is not the problem; it is how we use it that matters. Circumventing privacy by grouping personal data into “cohorts” and pretending we’re off the hook is a way to trick the letter of the law. However, more importantly, it is a way to trick users from the transparency and choice they deserve. And also, maybe, this is a way to trick the Open Web into using suboptimal methods to identify people and leverage marketing data to power advertising investments.
Another critical issue is that this may provide an opening for the large platforms to – in all impunity and under the disguise of (forced) consent — to exploit the largest “personal data vacuums” ever created.
The road to a privacy-first shared user identification future is a bumpy one. Using cohorts may seem like a safe way to avoid the bumps on the road. But they are merely a detour. And as often, taking a detour is longer and less efficient. This is particularly true if that road is controlled by an entity that can raise the toll price at will.