Zenith Digital Unplugged: Being future-ready with Data Clean Rooms

Authored by Rashmi Sehgal, Digital Transformation Lead, Zenith

The Journey for “Consumer Privacy Norms” started in 2017 and has been the main driver behind new industry adoption of data regulations, tracking limitations, and data access. Industry response has led to updates that have demonstrated impact across the ad technology ecosystem, from basic conversion limitations to advanced measurement and targeting restrictions. One such advanced measurement & data coloration effort has been towards “Data clean rooms.”

Firstly, let us understand what Clean Rooms are or what do they do. Clean Rooms are secure environments that facilitate data collaboration or the matching and analysis of data, with consumer privacy and data security at the foundation. They enable all this without exposing data thus are completely compliant with Data Regulations.

A clean room’s main purpose is to mitigate consumer privacy and data security risk while maintaining critical marketing and media use cases.

The core functions of a clean room include:

Protecting data custody (match data without moving data or sharing with 3rd parties)
Identity is not leaked (linkages of individuals cannot occur)
Attributes cannot be appended at user level (differential privacy)

Clean Rooms are continuing to develop as a privacy-enhancing technology solution, but the main goal of a clean room is data collaboration in a manner that mitigates risk while supporting critical marketing use cases. Several of these use cases are:

Incrementality Testing & Experimentation
Reach & Frequency
Attribution
Audience Insights & Overlap
Lookalike or Propensity Modeling
360 View of Customers
In Platform Conversion Optimization
AI-based Targeting

Currently, there are no standards for “clean rooms” in the marketplace. As such, everyone calls themselves a clean room. Each solution offers various levels of security, protections and privacy. Below are various categories of privacy enhancing technology that help facilitate data collaboration.

Safe Haven (LiveRamp, Epsilon, Experian)
Cloud Specific (AWS Bastion, Snowflake)
Closed Ecosystem (Google’s Ads Data Hub)
Boutique – Cloud Agnostic (Infosum, Karlsgate, Habu)

A brand will need multiple clean rooms to develop a holistic understanding of their customers. When it comes to identifying the clean room you may need to implement, it is situational depending on various factors:

Use Cases - what use cases are most impactful to your business?
Privacy & Security - how many threat vectors does the technology mitigate for these use cases?
Data Science – do you have advanced programming language in SQL, R, Python support?

The main goal of a clean room is data collaboration in a manner that mitigates risk while supporting critical marketing use cases. As you get started, evaluate vendors based on priority use cases, business objectives and how many thread vectors it mitigates, but keep in mind the general key criteria is:

Protect data custody (match without moving)
Identity cannot be reverse engineered or leaked
Attributes cannot be appended at user level (differential privacy)
Data security at rest