Answers to your burning clean room questions

What are clean rooms?  

Clean rooms are secure data collaboration platforms that allow marketers to responsibly share, combine, and leverage data sets without exposing directly identifiable personal information.  

What makes a clean room privacy-safe?  

Clean rooms are privacy-safe in a few ways. First, only your brand can access your space within a clean room unless you invite others to collaborate. If a brand uploads first-party data to clean rooms that don’t clean the data for privacy protection as a next step, there is still guarded access to that data within the clean room.  

Some clean rooms clean data after it is loaded, essentially removing elements of identity through hashing or pseudonymization. An example of this would be a clean room taking a customer’s email and changing it to a string of characters unique to the customer (otherwise known as hashing) but lacking any personal identifiers. Many brands would use this method on their first-party data anyway to be extra compliant.  Customer ID (in the retail realm) is an example of a privacy-safe way to attribute behaviors to a specific shopper without having their personally identifiable data.  

Additionally, most clean rooms aggregate data for extraction. While you might be able to see individual level data within the clean room, you cannot pull that data out of the clean room due to privacy regulation. What goes in does not in fact come out.  

So I need to know how to code to use a clean room?  

For the most part, yes. There are some pre-built queries within clean rooms such as Amazon Marketing Cloud that a non-coder could play around with, but the best insights will be extracted by data analysts who are experienced in commerce-related analytics. We have a whole team of those if you’re looking for help.  

What clean rooms should I pay attention to? Should I be using more than one?  

This depends on your business and even what area of the business you work on, as well as what exact data you’re wanting to look at. We would encourage brands to explore all oftheir data where it is available, which likely involves many clean rooms. Because of retailer walled gardens, brands find themselves often looking at first-party data in one clean room such as AWS or a non-retailer clean room such as LiveRamp, Amazon data (third-party) within AMC, and other retailer data where the partnerships exist.  

Brands will likely continue to find themselves moving data through different clean rooms as they’re able to in order to connect the dots, for example taking household identifiers from VideoAmp, attaching those identifiers to IDs within LiveRamp, and bringing those IDs into AMC to match with Amazon CIDs. Complicated, but exciting!  

Wouldn’t one consolidated clean room be a more efficient solution?  

Yes, but retailers will never allow brands to completely extract their data to be inputted elsewhere for comparison against other retailers. While looking at all your retailer data at the CID level would be amazing, this is a huge risk for retailers. Brands would better be able compare performance between retailers - leading to investments being pulled out of underperforming retail media networks.  

The profitability of a retail media network that brands are currently investing in is simply too high for retailers to be willing to let their data leave their ecosystems. We are, however, seeing a shift in how retailers are approaching building addressable audiences by allowing brands to look at their behavioral data or utilize behavioral data in their media strategies. More on this coming soon.  

What is it specifically that prevents me from putting retailer data into an external clean room?  

Taking Amazon Marketing Cloud as an example, within the AMC console, a data scientist can look at all the data that Amazon provides from organic purchase data to ad attributed data for all ad types and even user-level data via customer IDs.  

If the data scientist were to try and export the data, it would export in an aggregated format for privacy purposes (both for Amazon and Amazon shoppers) to the point where CID is no longer an available piece of data. The insights that AMC unlocks stay within the walls of AMC, and this is similar across retailer-owned clean rooms or retailer data partnerships (such as Walmart Connect x The Trade Desk).  

What makes AMC a critical place to be?  

We’re back to the value of customer ID. Amazon allowing advertisers to measure behavior at the user-level within AMC has paved the way for a shift in how shoppers are marketed to.  

We can now tie purchase data (SKU-level data) to behavioral data at the user level, creating a one-to-one ability to personalize messaging to shoppers at scale with precision like never before.

In addition to this level of behavioral insight combined with purchase data, AMC’s Audience-building capabilities allow us to group shoppers who demonstrate similar behaviors at each stage of the purchase journey together and market precisely to them with specific tactics to achieve a goal.  

Before all these pieces were accessible to advertisers, it was a guessing game to reach the right consumers at the right place, at the right time, with the right product.  

When will other retailers open clean rooms for brands to access?  

TBD! But our guess is soon. Amazon’s lead in this space only makes them stronger, and once advertisers see the level of insights that AMC provides, they’ll become hesitant to increase investments elsewhere without granular proof that those investments are worth it.  

We do see retailers and retail media networks moving toward sharing behavioral insights, though, even if not in the format of a clean room. Some of Instacart’s recently released tools aggregate behavioral insights on behalf of the advertiser for targeting, and we know that Walmart is also working to allow advertisers access to behavioral insights beyond the clean room partnership with The Trade Desk. The retailers are shifting their approaches from walled gardens to hedged gardens to accommodate demand from advertisers for better measurement capabilities as they try to verticalize ROI.

There’s no doubt that this is where the industry is heading, shifting toward behavioral insights and purchase data as the guiding forces for media strategies which can be fully measured in a deterministic manner.

‍