Retail Media’s Business Case for Data Clean Rooms Part 1: Your Data Assets and Permissions
It’s hard to have a conversation in adtech today without hearing the words, “retail media.” The retail media wave is in full force, piquing the interest of any company with a strong, first-party relationship with consumers. Companies are now understanding the value of their data and how that data can power a new, high-margin media business. The two-sided network that exists between retailers and their brands turns into a flywheel for growth.
As an emerging, or established, retail media network, it’s no doubt that you have considered a data clean room and its capabilities. While the retail media industry is set to grow by 25% year over year, surpassing $100 billion dollars by 2027, each retail media business needs a growth strategy that introduces differentiation to capture their fair share of growth from advertisers' increasingly limited dollars. At the end of the day, the top priority should always be your customers, and strategies that focus on secure data accessibility will have the greatest impact on delivering great experiences, relevant products and engaging content.
While retail media revenue is often determined by looking at a percent of total retail sales, it is not easy to capture incremental investment from brands that are already negotiating with your teams for shelf space, product placement and promotions. Retail media is alluring because it allows brands to target shoppers with the right message during the buying journey. Brands now have the ability to access otherwise inaccessible data, in a privacy-preserving manner to drive more performant advertising with closed-loop measurement to demonstrate the value of the investment. Increasingly, though, the managed-service model isn’t enough. Brands are demanding more — more transparency, more access to data — and extensibility beyond advertising, to use cases like product assortment, research and development, and inventory management.
The good news is, there are technologies to support this (enter: Data Clean Rooms). However, Data Clean Room integrations can be complex and often require an investment from both the business and IT organizations. Following a phased approach with clear commercial models to justify the investment can help remove some of the complexity and unlock more immediate value.
There are three steps to building and monetizing a data clean room business:
- Assessing your data assets
- Defining your data structures and permissions
- Unlocking revenue at scale with commercial models
Step 1: Assessing your Data Assets
Before venturing down the clean room path, it is critical that you assess your first-party data assets and resource availability. Some key questions to ask of your business include:
- How large is my first-party data asset? What percentage of my data is identified versus unidentified?
- How is my data organized? Is it easily accessible through a cloud or is it housed across multiple data warehouses?
- What technical resources do I have available to support implementation and ongoing product development?
- Do we have data quality and hygiene considerations? Are there any gaps in accuracy, completeness and timeliness that we need to address?
Your data clean room strategy will differ depending on your answers to these questions.
Questions | Answer | Recommended Next Steps |
Identified data asset scale? | Small | Determine how you can improve customer identity and grow your user base. |
Medium | Outline current retail media revenue by media channel, to define growth priorities. | |
Large | Ensure identified data is easily accessible and timely. Document any gaps. | |
Data accessibility? | Messy / fragmented | Start by centralizing key data assets for monetization. Build a single source of truth. Consider data-hygiene needs as you harmonize your data. |
Organized / accessible | Document what assets are available and begin use case mapping (segments-level vs. transaction-level data). | |
Technical resource availability? | Minimal | Look for partners who integrate with your existing tech stack and offer no-code options. |
Strong | Define priorities for when and how to leverage technical resources (implementation only vs. ongoing requirements). |
Once you have next steps identified for your business, it’s time to launch a request for information or request for proposal to understand which data clean room partners are best suited to meet your needs. Be sure to include questions in your RFI/RFP to understand:
- User requirements: What skill sets do platform users need to get to business outputs? This will help you assess technical resource needs to unlock value.
- Commercial models: What is the partner’s commercial model? Even if you’re not getting to pricing yet, this question will help you determine how clean room costs may impact your P&L. Clean room commercial models vary from a SaaS subscription model to a revenue sharing model to a compute model to pay-as-you-go model. Note, while the pay-as-you-go model can be a starting point, to build at scale, it’s not recommended, as it does not give you the financial leverage over time.
Step 2: Defining your Data Structures and Permissions
Now that you’ve started to identify the opportunity for your business and kicked off the partner assessment, it’s time to get your data ready for clean room monetization. This can be done simultaneously with the RFI/RFP.
The first requirement is defining the data structures and permissions that you will allow in your data clean room. Key questions here include:
- Will you allow user-level data to be available to partners? If so, how will you limit the risk of reidentification?
- Will you allow access to transaction-log data? If so, how will you ensure competitive brand data information is kept confidential?
Keep in mind: This will likely be an evolution. You might start with user-level audience data that can be compared to pre- and post-campaign impression logs, and later open up brand transaction-level data for full closed-loop analysis.
The key here is that the data you make accessible via clean room partnerships needs to be organized and clean, with intuitive headers and clearly defined use cases. Once you align on the permitted use cases, you can define the business models and revenue opportunities for your clean room application.
The primary use cases for clean room monetization in retail media include:
- Audience Segmentation: Allowing partners to build their own audiences using your first-party data, whether or not they bring their data into the audience (more on data collaboration as a use case later).
- Measurement and Attribution: Allowing partners to access your transactional data and media impression logs for measurement and attribution. This may include running predefined measurement models and/or allowing partners to bring or build their own models for analysis.
A note on publisher clean rooms:
Some publishers, like Amazon (via Amazon Marketing Cloud), Google (via Ads Data Hub), and Meta, provide access to user-level data at no charge, as long as you work within their data environment. There are two key considerations in pursuing a publisher clean room: (1) What are the trade-offs in sharing your data directly into a publisher’s ecosystem; and (2) Are you able to provide access for your partners to run their own queries? The monetization and commercial models outlined in the forthcoming Part 2 are based on allowing partners secure access to your data — in some cases with publisher data — to power these use cases. Publisher clean rooms will have a place in the post-cookie world, but they should be considered strategically in your roadmap.
- Insights: Insights is intentionally separated to designate use cases which may include separate data sets and non-advertising use cases that don’t involve traditional media measurement and analytics. Examples include basket analysis, product affinity, shopper insights, etc.
- Data Collaboration: Allowing partners to bring their own first-party data to the clean room for 1-on-1 or multiparty collaboration. In this article, we group Audience Overlaps into Data Collaboration because of the nature of joining multiple data sets.
A note on the Data Collaboration use case:
While your data is highly valuable in and of itself, in many cases your partners will ask to include their data in the clean room. This may be for a seemingly simple audience-overlap analysis or a more complex measurement use case. Keep in mind that joining two first-party data sets increases the complexity of the use case, including identity matching. Given the added complexity with data collaboration, we’ve separated it out as a high-value use case that may come later in your product journey, or may only be something you make available to top partners with highly sophisticated technical teams. Remember that the more you can reduce time to value, the stickier your product becomes.
With each of these use cases, the more granular data you make available, the greater the premium you can charge. For example, segment-level data may have the lowest price point, while transaction-level data summarized at a brand level or a regional level may command a higher price. Transaction-level data at a SKU-level is likely in the most demand and should be considered your premium product.
Your monetization opportunity will differ based on these inputs:
Topic | Answer | Premium |
Data Granularity | Segment | Low |
Aggregate / Sample | Medium | |
SKU-level | High | |
Use Cases | Audience Segmentation | Low |
Measurement / Attribution | Medium to High | |
Insights | Medium to High | |
Data Collaboration | High |
Once you’ve made it through Steps 1 and 2, you’re ready to start building a commercial strategy to define your market opportunity. Keep in mind that every business will look different based on your data assets, your permitted use cases and your partner selection. While you may not have made a data clean room partner selection yet, you’ve gathered information to better understand how the technology can fit into your P&L.
In our next post, we’ll go in depth on Data Clean Room Commercial Models for Retail Media, with detailed information on the three most common models for monetization.
—-----------------
To learn more about the Snowflake platform and our cloud-native DCR solution to help power your collaboration use cases, check out this recent blog post.