Cloud infrastructure technology services are converging on the media and marketing industry.
Snowflake has become a major programmatic player in the past year, while the Google Cloud Platform is deeply intertwined with its ad business through Ads Data Hub, its clean room and ad analytics hub.
The sleeping giant is Amazon Web Services. And AWS has woken up.
In July 2020, AWS hired Adam Solomon as global head of data collaboration and interoperability solutions for advertising and marketing. That’s a quite a mouthful. But he has a full plate to get to.
Solomon had stints as a digital ad leader at Time Inc. and Hearst and was chief product officer at PebblePost and chief growth officer of Lotame previous to the AWS role. So he knows all about the data and identity issues plaguing digital marketers and ad tech.
Recently, AWS launched its first clean room product, entering a crowded ring of independent vendors, social platforms and other data infrastructure tech competing for marketer mind share. And AWS is immediately the 400-pound gorilla in the category.
AdExchanger caught up with Solomon about the AWS clean room product and his vision for AWS ad tech and marketing services.
AdExchanger: What’s the main purpose for AWS clean rooms?
ADAM SOLOMON: AWS clean rooms make it easy and secure to analyze and collaborate on combined data sets across different customers without sharing the underlying data.
Advertising and marketing is our focus right now. But in the future, clean rooms could be used for financial services, healthcare, life science or other areas.
It could be a brand or a marketer that’s using AWS, or an agency. One announced customer is Omnicom’s Annalect. And they can collaborate with, potentially, a media platform, such as Dish Networks and Fox, which are launch customers. It could be a measurement company like Comscore.
Any combination of AWS customers can use clean rooms to create their own privacy-enhanced collaborations.
Is the Amazon Marketing Cloud clean room [which is an Amazon Advertising product, not AWS] built on different tech?
Amazon Ads announced that they are going to be migrating Amazon Marketing Cloud to the AWS clean room service. They are going to use the AWS clean room under the hood to power Amazon Marketing Cloud.
They developed their own clean room tech but decided it made more sense to migrate to the AWS clean room?
Correct.
We’ve heard from customers, Amazon Ads being one example, and other customers across advertising and marketing technology that built their own clean rooms. Building and maintaining clean rooms is actually quite challenging in terms of the query controls, the kinds of analysis controls, the audit logs and other things that are needed to scale.
Does it make more sense for the clean room tech to be a cloud infrastructure service?
Because AWS clean rooms are an AWS service, it means customers can build and integrate clean room capabilities into their own applications.
It’s not just marketers and publishers or agencies and publishers. There are customer data platforms, customer experience platforms, DSPs, SSPs and measurement companies, for instance, that can integrate AWS clean rooms as part of their offerings.
As a quick example, a customer data platform helps marketers or media companies collect data from different touchpoints, then segment, analyze and activate. But perhaps a company has another data repository in an ecommerce system, a subscription system or some other data source that the CDP is not the system of record for.
The CDP would love to offer the ability to perform analysis across these data sets. But they don’t have control over one or more of those data sets because their client keeps it somewhere else, for good reason. A service like AWS clean rooms can be used by the CDP in partnership with that client to register data sets that live elsewhere in AWS, and to use that data without it ever leaving their system.
Are you working with any third-party clean room tech companies at launch?
There are definitely companies that have clean room capabilities, or even as their primary business, using AWS services in various ways.
We had LiveRamp as a partner for AWS clean rooms, for instance. There are some posts by clean room tech companies about how they may use AWS clean rooms.
You mentioned clean rooms as one of the privacy-enhancing technologies for AWS within advertising and marketing. What else is there?
There’s a robust conversation around different approaches and techniques for performing privacy-enhanced collaboration. Certain companies are looking at multiparty computation; some use technologies called “trusted execution environments.”
I mentioned before that a feature of AWS clean rooms is privacy-enhanced analysis on inquiry data. Doing that kind of analysis can be expensive to process using encrypted data. That’s an area where, if a customer is doing multiparty computation, say, AWS can really help, since we have specialized processors for different workloads or processing needs.
We’re also really focused on collective needs when it comes to media planning, activation and measurement. And we’re responding to those needs with features such as cryptographic computing. Cryptographic computing allows customers who use AWS to encrypt their data at the source, so the data is encrypted at all times. That’s something we heard directly from customers with certain sensitivities, where they wanted customer data encrypted at the source if it was to be used for clean room analysis.
What do you think are opportunities for AWS advertising and marketing services in 2023?
Data collaboration, an important part of my role, is also the notion of interoperability and identity resolution.
In 2023, AWS is going to make further announcements in terms of how we can help customers with identity resolution workflow. We’ve already made some announcements with LiveRamp and with The Trade Desk for Unified ID 2.0, where we’ve worked with those customers to reimagine how identity resolution operates in the cloud.
In the case of LiveRamp, they developed a product that we helped bring to AWS Marketplace called their embedded transcoder, which allows customers to transcode from one RampID to another. Every LiveRamp customer has their own encoding of a RampID. A marketer may have RampID 1234, but a publisher has 5678. So LiveRamp has a process called transcoding, which goes from one ID to another, and LiveRamp developed a version of this transcoder that runs in an AWS customer’s own cloud environment.
Similarly, with The Trade Desk for UID 2.0, we worked with them so that IDs can be deployed in a secure container inside a customer’s environment, to take email addresses and turn them into UID 2.0 IDs.
In addition to clean rooms, identity resolution is a challenge for industry customers. And that’s something we’re working on as well.
