January 25, 2023
Amazon Ads Debuts Ad-Targeting Solution Using AI, First-Party Data At Cannes
- by Laurie Sullivan @lauriesullivan, June 18, 2024
Amazon Ads has developed Ad Relevance to offset the elimination of browser cookies and inability to rely on one piece of information or IDs to accurately target ads.
The technology, publicly debuted at Cannes Lions Festival of Creativity, uses artificial intelligence (AI) to analyze billions of browsing, buying, and streaming signals tied to real-time information about content being viewed.
It understands where customers are in their shopping journey and serves them relevant ads across devices, channels, and content types without needing third-party cookies.
AI helps the platform to understand what creates great shopping experiences, and the connection those shopping experiences have with ad interactions, product and category interests, and cadence of actions along the path-to-conversion. And, it’s available on Amazon’s demand-side platform (DSP)
advertisement
advertisement
The technology already powers many of Amazon Ads offerings, including Amazon Audiences, Contextual Targeting, and Performance+. It supports cookie deprecation, which is still unresolved, and does not reply on third-party identifiers.
Advertisers already working with Amazon to use the technology has found it deliver extend addressability on up to 65% of previously anonymous impressions lowering CPMs by up to 34%, and improving CPCs by 8.8%, delivering on 100% of the budget.
While it’s important to simplify programmatic media buying, it shouldn’t cost advertisers control and transparency.
Amazon built reporting for the platform and has simplified fundamental marketing tasks like planning, activation, and measurement.
Performance+ digs into the data to provide advertisers with insights from Ad Relevance's programmatic media buying. Advertisers set key performance indicators (KPI) they want to achieve through advanced AI and machine learning models.
The reporting platform also provides transparent data for creatives, placements, audiences reached, and controls to ensure ads reach the audience they want to reach.
Advertisers upload business signals through Amazon Ads ad tag and conversion API. Performance+’s advanced AI and machine learning models use this information to automatically optimize campaigns.
Cox, a telecommunications company, which used Performance+ to drive more website visitors based on their online/call center engagement signals, leading to a 40% CPA improvement, the company said.
Laurie, what is usualy missing in these discussions about the benefits of targeting---which, in theory, are great---is how exactly they would be applied. For example, whenever I buy something on Amazon, I get numerous invitations to buy the same product again. Is that trageting? Also, if you are a headache remedy brand, and Amazon is promising to help you target your ad impressions, what do you do---go after those Amazon customers who have bought you in the pat---or those who have bought rival brands----or those who use similar products---or all three? And what if Amazon only accounts for 5% of your product category's total sales? How can it help you in targeting those of its many customers who might be receptive to your ads----even if they don't buy your prosuct from Amazon? etc, etc.
My point is not to bash targeting as a concept. But often it's not the all encomapssing solution it's played up to be. As a result, many "targeted buys on digital media are being made the same old way---via profiling by sex, age, region, etc.---where it is assumed that any consumer who falls into the desired demo break is a brand's prospect---which we know isn't true.