Amazon is working with top media holding companies and brands to make its data more available for use in their media planning, according to people familiar with the plans.
The e-commerce giant has been huddling with the agency world—companies like Omnicom, WPP, Dentsu Aegis and others—about how they can partner on the future of advertising on the platform, especially when it comes to applying data to targeting ads and measuring how those ads perform.
"Amazon, for the first time ever, is starting to realize that monetizing the data they have and making it available for purchase, not personally identifiable information, could open a revenue steam that wasn't there before," said one agency executive who is familiar with the talks Amazon is having with agencies.
Amazon is developing data and analytics tools for brands, backed by machine learning and its ubiquitous web services. Amazon already is running experiments with different agencies and brands, some that look at targeting ads and some that measure attribution, showing which ads lead to business results. There is a flurry of ongoing trials, but advertisers say they expect it will eventually lead to a coherent concrete data service operated by Amazon.
Agencies and brands will be able to build customized ad bidders and reporting tools, buy ad inventory across the web, and import consumer data to learn more about them, and ultimately build ad targeting models that are more exact. "The Amazon cloud already has a bunch of utility built into it," said a second agency executive who has worked closely with the technology.
The possibilities are greater than anything Amazon has offered so far through its first-generation self-serve ad platform or basic audience matching, advertisers say.
Amazon's secret weapon in its advertising push against the Google and Facebook duopoly is Amazon Web Services, the cloud computing platform used by companies as varied as Kellogg's, Comcast and Major League Baseball. Amazon Web Services has become one of the most significant technology layers undergirding industries around the world, and it is a key component to how Amazon will meld data and marketing, according to multiple advertisers.
To that end, Amazon's ad team has been promoting what's known as a "clean room" for complex data and analytics research, according to these advertisers, who spoke on condition of anonymity because they were not allowed to discuss details of their dealings with the company. "Clean room" is a generic name for a data-sharing platform that adheres to strict guidelines around privacy and tries to prevent any information from leaking.
That could also lead to more sophisticated data and analytics tools similar to Google's Ads Data Hub, which is a platform for advanced research into consumer behavior and brands' businesses.
"What Amazon is building will enable brands to craft a full-journey, attributable marketing experience," said Chris Apostle, the evp and head of performance at Havas Media, who actively leads the relationship between the agency and Amazon, but said he can't share further details on all of the retailer's data ambitions as they are still evolving. However, he has heard the term "clean room" and knows what direction that will take Amazon's advertisers.
"The digital data 'clean room' will provide [insights] into behavior across consumers' purchase paths," Apostle said. "This is very different than anything advertisers have been able to do with Amazon until now."
Amazon wants to use data as a lure for big advertising spenders to commit to investing in its platform, where advertising is the fastest-growing segment of the business. Amazon is expected to hit $4.6 billion in ad revenue in the U.S. this year, making it the third-largest digital advertising platform behind Facebook and Google.
Amazon already offers ad targeting technology, the kind that most digital ad rivals provide, like being able to their match customer e-mail lists to people shopping on Amazon to target ads to them. Amazon also has basic demographic and shopping data that let advertisers target consumers based on characteristics like age, location, gender and purchase history.
Amazon ad formats include search ads, which can be targeted to the queries people type into the search box. It also has display and video ads, which are not as easy to target because they don't come with an immediate signal like search intent. Google solved this problem on YouTube, for instance, by targeting video ads based on a viewer's search history.
Amazon is building a marketing ecosystem that could rival Google, though. It also has video, through Fire TV and Twitch, the streaming service similar to YouTube. Amazon also owns IMDB, which shows video ads, and also has publisher services that deal with websites just like Google's publisher ad network, where it can target ads outside the websites it directly owns.
What Amazon is missing is better technology and data to make sense of its sprawling marketing platform and unite its disparate formats. Advertisers say the design of its self-serve ad management platform is outdated and clunky. That's partly the reason why the company has worked so closely with the agency world and third-party marketing tech platforms like Kenshoo in the past year, to help test and design new ways of buying ads on Amazon.
Amazon's ad business was streamlined this year, too. For years, Amazon's ad offerings were a jumble of services with different acronyms and run by separate teams. There used to be Amazon Marketing Services known as AMS, and Amazon Media Group known as AMG, and Amazon Advertising Platform known as AAP. Now, all those fall under Amazon Advertising.
Amazon declined to comment for this story, and the exact details of any planned data products are still in the works. Advertisers said they were still in discussions with Amazon about how they could help develop these next-generation data services, and anything concrete would likely come next year.
What is certain is that sophisticated data partnerships are already forming. For instance, agencies have been using Amazon Web Services to analyze data from Facebook ad campaigns, according to a person familiar with the offering. That means advertisers are able to import data out of Facebook and dissect it in Amazon's environment, which requires a partnership between the two web rivals.
Amazon declined to comment on the arrangement. However, advertisers said that the whole point of the "clean room" technology is so that no data from Facebook leak into Amazon.
"It could get tricky for a lot of brands," said Andy LaFond, executive media director at R/GA Chicago, who is not involved in Amazon's plans, but is familiar with the platform. "Data is Amazon's power for sure and brands have to be really careful about what data they're willing to share with Amazon."
Amazon is already known for being able to give brands solid intelligence about ads on its properties and tell them when those ads lead to sales on Amazon. But the company is expanding its ambitions and wants to be able to help advertisers serve ads anywhere online and measure the impact, even if the final sale doesn't take place on Amazon.
Earlier this year, Amazon developed a pixel, a common technology that marketers use to measure ad performance. In this case, Amazon wanted to show marketers that its ads are more efficient than its rivals, and brands could attribute the results of marketing campaigns to their spending on Amazon.
Amazon's data ambitions include giving brands the ability to target their exact consumers, identifying what they bought down to the last nail, according to advertising executives. Also, instead of just targeting people by keyword when they're searching Amazon, the same type of targeting could be applied to display and video ads across the web, the advertising executives said.
"This is a faster, more direct way to get to data," said a digital agency advertising executive who works closely with Amazon. "From the search side it's easy, but from the display side it's a little harder to target that exact consumer who bought that exact product."
Amazon data products could also provide unique insights about consumers who buy certain products, like figuring out what else they might buy, even when the link is not so obvious.
"Amazon is dealing with the most valuable data asset in the world, even more than Google's search data," said an executive at a top marketing technology and analytics firm. "If I want to know what someone will buy, then I have to know what they already bought in the past."