
Ad tech providers are developing a multitude of identity solutions to replace third-party cookies. There are five main strategies for ad targeting in a cookieless world:
- Publisher first-party IDs
- PII-based deterministic IDs (LiveRamp, TTD)
- Probabilistic IDs (Experian)
- Data clean rooms (InfoSum)
- Device-based data (Google, Apple, ID Ward)
But being ‘cookieless’ does not mean that these solutions will work in a privacy-first world. To be future-proof these solutions will have to pass at least four more tests:
- Meet legal requirements around privacy (which I discussed in this post).
- Provide high-quality data: this means accurate cross-site data, and ideally cross-device data, about consumer interests and purchase intent
- Provide reach and scalability
- Be compatible with Google and Apple platforms
Description |
Privacy |
Reach |
|
Traditional DMP |
DMPs using 3rd party cookies and advertising IDs to get omni-channel view of the customer (e.g. BlueKai) |
Non-compliant: data is transacted without consent or control by the data subject |
Omni-channel reach, both online and offline |
Universal ID |
New products using deterministic IDs (e.g. hashed emails) to create an identity graph across the ecosystem (e.g. Unified ID 2.0) |
Non-compliant: companies can still match personal data using the universal ID, data subject rights are unenforceable and data transfer to non-EU countries cannot be prevented |
Cross-domain and cross-device reach using forced authentication; potential for offline matching with authenticated users |
1st-party DMP |
DMP processing first-party data from the publisher to create segments, also on the edge (e.g. Permutive) |
Compliant with caveats: compliance achieved only if DMPs do not overlay third-party data and if the user ID is not passed on to brands |
Without third-party data, segments can only be created with data from a single domain |
Publisher platforms |
1st party data platforms owned and controlled by the publisher/group of publishers (e.g. Ozone Project) |
Compliant with caveats: compliance possible only if publishers do not pass a user ID to vendors and brands |
Without third-party data, segments can only be created with data from a single domain |
Data clean rooms |
Next-gen data clean rooms connect multiple 1st party data sources without sharing PII (e.g. InfoSum) |
Compliant with caveats: data owners need explicit consent each time their data is connected with another company, there must be no way to de-anonymise individuals, no matching can be done with companies outside the EU |
Limited reach, one-to-one database matching between companies |
0 Comments