As 2021 approaches, there are three major changes in the new year that will disrupt the portion of the digital ecosystem that leverages audience data for various applications such as digital marketing, identity verification and personalization.
- First are the additional global privacy regulations, similar to GDPR, that are being introduced. As Gartner has forecasted, more than 60% of the world’s population will soon be under a GDPR-like privacy law.
- Secondly, Apple’s delayed IDFA policy change will come into effect, which will significantly impact the reach and scale of digital tracking in the Apple ecosystem.
- And lastly, the sunsetting of 3rd party cookies will continue its downward spiral as Chrome continues to make product enhancements that make them less and less relevant, until its full deprecation in 2022.
Technologists have been anticipating these changes for years and have developed a few approaches in preparing the industry for these changes with 3rd party cookies, Apple’s IDFA and additional privacy regulations.
One of the approaches leverages registration data such as email address to link individuals between brands and publishers. This is currently the most commonly adopted approach and typically involves storing and hashing someone’s email address. When someone logs into a brands’ website and makes a purchase, then that same person logs into a news site to view a piece of content using the same email address, marketers use that common email address to personalize and deliver relevant advertising or messaging. However, this approach has stirred the nerves of major ecosystems including Apple, who has publicly denounced such an approach. As a result, there is uncertainty around how well this approach will scale and be effective.
The second approach is what Chrome has shared in its Privacy Sandbox tool with the concept of “Federated Learning of Cohorts (FloC)”. FloC uses machine learning to develop cohorts, or individuals with common attributes, based upon site visits and it’s run on Chrome so data is completely distributed and privacy is completely protected since there is no information being collected or centralized at the individual level. Differential privacy can also be enabled to further distance the target from the source or masking the individual from being tracked or identified. The elegance of FloC is that the cohorts are determined within the browser, thus upholding a true decentralization. However, it also presents challenges for adoption. To start, the categorization of interests is relatively static, thus eliminating a lot of the flexibility of digital personalization in today’s demanding ecommerce environment. In addition, the portability of FloC is non-existent because it will be only available on Chrome browsers. This lack of portability should be recognized for what it is, an inhibitor to continuity in the consumer experience and to differentiation in channel optimization strategies. And lastly, while the specifications are still developing, the implementation of FloC is not trivial and will require a deep understanding of population biases and how the degradation of interest behavior applies to marketing campaigns.
A third approach in development is the “data bunker” approach that creates the first iteration of data decentralization. Each data bunker operates with a bespoke data owner and any consumption of that data has to operate within the confines of the data owner’s environment, thus the entirety of the data collection and matching are tailored to the rules of that data bunker. The advantage of such an approach is that data owners are tracked and managed, thus data leakage can be reduced while data rights management and usage policies can be enforced throughout the collection process. However, the downside is that these bunkers will create multiple persona shadows with little identity harmonization of customer data. This often makes downstream analytics challenging and disconnected. In addition, even though it’s decentralized to the bunkers, only a limited amount of data leaves the data owners’ firewall, thus it significantly reduces the amount that can be shared, resulting in limited match and reach capabilities of these bunkers.
How Adara Privacy Token Improves Upon Other Data Options
Adara Privacy Token has been developed to deliver upon the expanding use cases requiring a client’s data to NOT leave their firewall, which requires making the technology that matches and segments portable, so that it can be deployed within the client’s firewall. Clients have full control over what data is deployed for matching, and the use cases for which they want to match. Leveraging machine learning techniques and portable data onboarding SDKs, Adara Privacy Token delivers a pure, privacy-safe solution to clients who want to share and connect their data. The strength of this approach is that clients’ data never leaves their firewall, so they can match with a much broader set of data while at the same time, have full control over the precision in which they want to calibrate. In addition, clients have the full custodianship of their data and eliminate any concerns of leakage, breaches, and privacy exposures. Because Adara Privacy Token provides for using a broader breadth of data for matching, it powers significantly better matching precision and recall than the traditional registration data approach using identifiers such as hashed email.
At Adara, we fundamentally believe that protecting digital privacy is the only way of developing a meaningful digital strategy. And to protect and ensure digital privacy, brands need to have full custodianship of their data with the right controls and techniques for de-identification and matching. We are continuing to explore better ways of serving the needs of our customers, and Adara Privacy Token is a privacy-first approach to enable a future-proof digital strategy.