Amazon unveils Ads Relevance, claiming it is no longer reliant on single IDs

Amazon’s Ad Relevance is an offering that claims to help advertisers target online audiences without the need for the (soon-to-be-departed) third-party cookie or reliance on the host of ID solutions that have flooded the market in recent years within its DSP.   

“We have essentially, over the past three years, rebuilt our system based on machine learning models,” Tomasette told Digiday, claiming that its ad targeting and measurement system is “no longer reliant” on identifiers such as third-party cookies.  

Amazon’s Ad Relevance, which was tested throughout 2023, uses signals such as browsing, shopping and viewing behavior across Amazon properties to predict user behavior and then offers media buyers ad opportunities. 

“We think the future is using machine learning and model-based audiences … we think ad-identifiers have a place in that,” Tomasette said, adding that Amazon’s system is “no longer solely reliable on advertising identifiers.” 

Tomasette explained that in its current guise, Amazon’s advertising platform does use ad identifiers such as third-party cookies, etc., but the notion of replacing one identifier with an alternative one was limited.

“We don’t think that’s a long-term solution,” he said. “So, we’re not betting on our deep understanding of the shopping experience and how we built machine learning and AI models, and that’s expressed in specific options in our DSP.”   

In many ways, Amazon’s strategy emulates that of other Big Tech companies, such as Google, with its Performance Max offering. The retail giant has made much of its Performance+ offering in recent weeks, claiming that it works by using “first-party signals and machine learning to automate campaign setup, audience creation and optimization,” according to documentation recently obtained by Digiday. 

Nathan Woodman, founder of media consultancy Proof in Data, told Digiday that Amazon’s key differentiator in the market is the amount of purchase signal behavior at its disposal.

“The key things are about addressability and relevancy,” he said. “The industry often confuses the two; addressability is about finding the right audience, whereas relevancy is about understanding the context that person might be in, and handing them the right creative.”

Related Insights

Member Exclusive

Pitch deck: How Amazon is talking to marketers about Performance+ Subscribe To Read Read More

He added, “You don’t even need to know who the person is in order to hand them the right creative … you can do this using the machine learning algorithm to to piece together a multidimensional cohort.”

Amazon Ads’ latest announcement follows separate developments with media agency holding groups Omnicom Media Group and WPP’s GroupM that emerged in a week that many consider the zenith of the advertising industry’s dealmaking.

The partnership with OMG lets staffers there to access Amazon’s proprietary browsing, shopping and streaming insights to directly tie linear and CTV investment to purchases made on Amazon.

Meanwhile, its GroupM tie-up enables clients at the WPP entity to develop “shoppable content” on Amazon Live, the free ad-supported FAST channel on Amazon’s premium streaming services, such as Prime Video.

Amazon Ads’ revenue is expected to grow approximately 25% in the U.S. this year, according to forecasts from eMarketer, to nearly $42 billion, with analysts there estimating that its market share will be near 14%, behind only Meta and Google.