Unlocking Privacy-Safe Customer Analytics and Data Collaboration
By Tim Burke, Affinio CEO
Privacy-safe data collaboration is a hot topic. With a new focus on privacy, the marketing & advertising ecosystems are experiencing massive disruption with changes like FLoC & ATT. These changes are forcing a paradigm shift that will require a scalable, privacy-safe data collaboration and analytics solution across brands, publishers, agencies and data providers.
Applications for Privacy-safe Analytics and Data Collaboration
Most Snowflake partners require data to be extracted before being analyzed. The results from this external processing then get returned to the data warehouse. However, with every data extraction from the warehouse comes increased privacy risk. In addition, many of these analyses are repeatable workloads from marketing and advertising teams that should be “productized” and pushed down to the data layer for privacy, security and speed.
To address this challenge, we’re excited to announce the release of our first multi-party analytics applications that run natively inside Snowflake's Data Warehouse, requiring no data movement. Unlike other solutions that require you to onboard PII to a 3rd party broker or require sharing private data into vendor-controlled “bunkers”, we give you full control of your private data and never require it to be shared nor exposed to us. Instead, we act as a service provider enabling multi-party data collaboration between partners so you, the data owner, maintain 100% control and your private data never leaves your Snowflake environment.
Our applications are designed for scale and solve for repeatable workloads with all analytics running inside the data warehouse. Our initial release includes the following applications:
- Customer Profiling: unlock customer analytics and insights across all data in the warehouse.
- Propensity Modeling: automatically score all your customers on their likelihood to churn, upsell, engage, etc.
- Segment Overlap: identify ID overlaps between your customer data and multiple other strategic publishers and data partners simultaneously in one view (without ANY party exposing private data or PII).
- Look-alikes: automatically create scalable, custom audiences, based on first party data “seeds”, within strategic publisher’s / data partner’s Snowflake environments ready for activation.
How Do They Work?
Affinio’s applications are created using Snowflake’s Secure Functions and leveraging Snowflake’s Data Sharing capability. We create these applications inside Affinio’s Snowflake instance and can pass them directly into a customer’s Snowflake instance using a Data Share (no data is shared, only the applications). The customer can then execute these applications, run customer analytics, and conduct privacy-safe, multi-party data collaboration with marketing and advertising partners all from their own Snowflake Data Warehouse.
Does it mean changing my tech stack?
Since our applications run inside the data warehouse, our technology and analytics can be leveraged by any other 3rd party vendor or platform that connects to Snowflake. As such, our technology augments and extends the capabilities of the existing tech stack - without requiring any “rip-and-replace”, minimizing change management challenges.
Imagine you’re a B2C brand, with customer first-party data stored in Snowflake, and are looking to have your agency help acquire more high-value customers but don’t want to share or expose your private data to anyone. By leveraging Affinio’s applications, without needing to share or expose your row-level PII, you can enable your agency to analyze your customer data in aggregate, develop data-driven strategies, identify the best media channels to reach more high-value customers based on overlapping segments, and create scalable look-alike audiences for media activation. All this without moving data.
In the coming weeks, I’ll be providing detailed examples of how our applications can be used in unique configurations between brands, publishers, agencies and data providers to unlock scalable privacy-safe data collaboration.
This is just the beginning - our roadmap will continue to bring more workloads into the data warehouse unlocking new opportunities for advanced data collaboration across the ecosystem. We’re interested in working with partners to help guide our product vision -- if you’re interested, or have ideas for applications you’d like to have developed, please sign up below.