Data Clean Rooms

An effective data-driven solution for modern advertising

Intro

Unless your head has been buried in the sand for the last few years you already know that third-party cookies and ID deprecation is coming at us quickly. A tighter privacy framework and growing Walled Gardens have us all scratching our heads for technical solutions that allow us to develop data-driven marketing strategies and top-notch Customer Journeys.

The Marketing Data Technology (MarDaTech) space is hungry for compliant solutions that allow relevant advertising, accurate campaign performance measurement and media attribution. Everyone is looking for safe mechanisms to understand prospects and customers as well as to identify them across different online environments.

At Bedrock we have started to help our clients navigate this changing and uncertain situation with effective, yet compliant solutions. One of which is Data Clean Rooms (DCRs).

DCRs deliver security, privacy, and data governance controls and processes. Moreover, data science can perform queries to build an analysis layer on top of them.

Data Science on Data Clean Rooms

These neutral environments offer statistical rigour, analytical possibilities and privacy protection. In data engineering jargon, DCRs are Data Warehouses where brands and/or advertisers exchange, can share and enrich their own data to run and measure optimised advertising actions. DCRs enable two or more parties to bring data together for joint analysis and custom data science developments.

For data sets (e.g. from CRM or transactions log) to be loaded, PII data needs to be “hashed” for transmission and uploaded, and then appropriately secured and encrypted. The resulting datasets are merged into aggregates and later divided into cohorts (as groups of users). This way, marketing and data science teams can work together to find value across these cohorts without relying on user-level data or individual addressability.

Aside from advertising, data scientists can apply their analytical techniques and algorithms to find insights across these groups, including; common relevant characteristics of high-value clients, their passions, interests, unexploited user profiles and even predict cohorts’ lifetime value.

By performing multivariate statistical analysis, the most important data attributes that allow us to understand the clients can be derived. Through PCA (Principal Component Analysis), MDS (Multi-dimensional scaling) or density-based unsupervised clustering, we can understand the taxonomies for each client base; and to understand the evolutions of these groups we could rely on algorithms such as STATIS methods.

Aside from advertising, data scientists can apply their analytical techniques and algorithms to find insights across these groups.

Three groups of companies are offering them:

1. Walled Gardens : Google or Amazon. A large number of companies using Data Clean Rooms are those that have already spent and still invest a significant chunk of their advertising budget within the Walled Gardens. In return, the big players allow brands to obtain user-level data, though this is only available within those environments. There are two main Data Clean Rooms provided by the Big Tech as Facebook Advanced Analytics (FAA) was deprecated as of July 1st 2021. Google Ads Data Hubs (ADH) was one of the first Data Clean Room solutions in 2017 and it was developed as a privacy-based replacement for their previous advertising solution called DoubleClick. It is a privacy-native warehouse built on Google Cloud, and Google BigQuery is where we construct and perform queries that get our clients the insights they are looking for. No personally identifiable information (PII) is stored in BigQuery, which is excellent for processing large data volumes coming from Google owned tools and data sources such as Campaign Manager, Display & Video 360 (DV360), Google Ads, and YouTube. Amazon Marketing Cloud (AMC) is built on the Amazon Web Services cloud, this solution helps our clients discover the net impact of cross-channel media investments, perform analytics across multiple pseudonymised data sets, and generate aggregate reports. A unique feature is that anyone can use Amazon Marketing Cloud without having to have an AWS account.
2. Emerging tech companies like InfoSum, LiveRamp (Safe Haven) Optable, Habu (CleanML), Permutive. These companies run independently and act as a middle man between two companies who want to exchange data. They build neutral rooms on demand for those that seek to adapt their advertising strategy so that a brand and a retailer can match data and refine audience segments, for instance. All of them offer pre-built turnkey integrations to ease data collaboration and they work across many verticals, including retail, CPG, travel, etc. In most cases only two parties are allowed at a time, but there are exceptions, i.e. Snowflake (distributed data clean room) enables data sharing with multiple parties at the same time, while keeping the same level of security.
3. Organisations that own huge amounts of users and content data(e.g. membership platforms or retailers). Some business models related to long-term subscriptions and memberships tend to own and acquire large amounts of visitors’ and users’ information, such as Netflix or TikTok, allowing them to construct their own instances and convince providers and/or other retailers to partner and exchange their own first-party data. An example is Disney’s Advertising Sales group that launched their own clean room solution last October 2021 using a combination of Snowflake, Habu, and InfoSum tech stack. Now Disney allows advertisers to access thousands of first-party segments from its vast portfolio of brands across media platforms.

What to do next?

At Bedrock we help organisations to decide which Data Clean Room software fits them better as we already work with all major providers of Cloud infrastructure. We not only support the implementation of DCRs, but we can also help you integrate new data sources and provide data intelligence and analysis on demand.

Reach out to learn how we can help you leverage best-of-breed Data Clean Room technologies and value-added services!