Thor Olavsrud
Senior Writer

Disney data clean room helps advertisers leverage audience analysis

Feb 14, 2022
AnalyticsDigital Transformation

Aiming to give advertisers access to a vast quantity of audience data while protecting users’ privacy, Disney Advertising Sales has taken a data clean room approach to data governance.

big data analytics analysis thinkstock 673266772 100749739 orig
Credit: Huawei

For years, Disney Advertising Sales has been staking out the new frontier in data and analytics to better help its customers find the right audiences for their messages.

The explosion of The Walt Disney Co.’s streaming services has added massive amounts of data to the mix. Disney Advertising Sales, responsible for advertising sales and integrated marketing for The Walt Disney Co.’s entertainment and sports offerings, can now provide advertisers access to more than 1,000 user segments built from that data. But with great power comes great responsibility. Disney has had to reimagine its approach to data governance to ensure it protects its users’ data and privacy.

In October 2021, Disney Advertising Sales unveiled a new clean room data solution built with the help of vendors Habu, InfoSum, and Snowflake, among others. Data clean rooms are places for partners to bring data together for joint analysis under defined restrictions.

Dana McGraw, vice president of audience modeling and data science at Disney Advertising Sales, says Disney’s relationship with its guests is the guiding light for everything the company does with data.

“That relationship with our guests is why our ad offering is so compelling, because of our content and because of the way that we relate to our guests and the way that they relate to us,” McGraw says. “As we think about data, data use, data governance, it’s really all about, ‘Does this improve the experience for our guests?’”

Data clean rooms allow sharing data safely

“A clean room solution is a way to allow brands to access insights about their own audiences, and who they want to advertise to with us, without any kind of exchange of data between us,” McGraw adds.

Snowflake’s Data Cloud is a cornerstone of the clean room solution. Its data-sharing technology, private data exchange platform, and secure function and secure join capabilities enable the clean room solution.

“The exciting opportunity with the Snowflake Data Cloud is that it provides all of the safety protection that we have from a data standpoint and allows us to do really interesting things with audience graphs and other datasets, client data, other third-party datasets that we believe the marketplace is very hungry for as it relates to insights, activation, and measurement,” says Lisa Valentino, executive vice president of client solutions and addressable enablement at Disney Advertising Sales. “The Snowflake solution allows us to manipulate data at scale, in an environment where we feel very comfortable.”

Valentino explains that many of Disney’s clients are looking at clean rooms to derive pre-planning insights by connecting their own first-party data with Disney’s data. Pre-planning insights are critical to “upfronts,” gatherings of television network executives, major advertisers, and media at the start of important advertising sales periods that allow marketers to buy commercial time “up front.”

Over the next several months, Valentino says Disney plans to share insights and best practices about working with its data in the clean room. She hopes that will help clients better understand how Disney’s data adds incremental value on top of their own data, and how to architect their data so it can sit inside the clean room.

The Snowflake Data Cloud enables Disney to have a “single source of truth” regarding its data, while keeping it secure, available, compliant, and easily accessible for its partners. That single copy of data also gives Disney scalability and flexibility in how it prioritizes workloads, allowing it to better support its BI, analytics, data science, and machine learning teams while minimizing the time data engineers must spend orchestrating, organizing, and building out data pipelines to deliver data from various sources. The clean room gives Disney the ability to define whether, how, and when to provide query results to ensure data remains anonymized and secure.

Disney Select powers the data clean room

Behind the clean room is Disney Select, which pulls together all of Disney’s first-party data and advanced modeling capabilities under one umbrella. Disney Select, in turn, is built on the Disney Ad Sales Audience Graph, which is designed to map relevant and available IDs across the Disney Platform for a particular household and connect attributes and engagement across all Disney’s endpoints. McGraw notes that Disney Select gives marketers the ability to choose their desired audiences from a library of more than 1,000 first-party behavioral and psychographic segments.

“We’re operating with over 100,000 attributes to inform these audience segments,” McGraw says. “We’re leveraging advanced machine learning, so we’re able to do a great deal of modeling, whether off of a seed of information or off of appending third-party data.”

By way of example, McGraw says Disney might not have a lot of internal data about auto buying, but it could append an automotive marketer’s data to Disney’s data in the clean room to create a more tailored model.

“Around each category, we’re thinking about the desired outcomes and then we’re doing modeling against that to create segments for the desired outcome,” McGraw says.

Embedded data science has been the key

Disney has taken an embedded approach to data science to drive success in this field. McGraw’s data science team sits inside the business in Valentino’s department.

“We’ve seen great success in integrating this solution team inside of our go-to-market team,” Valentino says.

McGraw adds that thriving as data scientists within an ad sales organization has required being thoughtful about talent and ensuring a diversity of background and skillset within the team. It’s not enough to strictly hire members with quantitative backgrounds.

“We want to have people with a marketing background sitting right next to those with heavy quant skills,” McGraw says. “There’s an exchange of ideas and understanding and workflow across those groups. Whether it’s data science, advanced analytics, or data solutions and activation, those groups exchange ideas, they exchange skills, and they work together hand-in-hand.”