2026 may be the year of AI, but the buzzword as television advertisers head into the upfronts season is “outcomes.”
And Amazon Ads is drooling about the sales prospects.
“TV is historically largely reach-based, and most upfront spending still reflects that,” Lily Tong, Amazon Ads director of brand and cross-channel measurement, told AdExchanger. “But what we’re seeing from our side is a clear shift, not just reach but also audiences, and increasingly toward outcomes.”
To that end, Amazon has packaged a handful of upgrades to its ads measurement solutions, obviously catered to TV and streaming media advertisers.
For example, last November, Amazon Ads launched a beta program for what it calls Prime Video Insights, which are being opened up and expanded with more data sets. The tool allows advertisers to combine their first-party data to see, for instance, which genres or types of shows or audiences led to sales. Typically, an open programmatic CTV advertiser has next to no visibility into the programs or movies they’re interrupting.
There is a new metric for modeling long-term sales values, which, Tong said, is “based on a heuristic with the brand’s own data” to quantify the long-term value from new-to-brand customers. Amazon Ads will also increase the measurement lookback windows available for cloud analytics.
The “backbone” behind these changes, according to Tong, is the “unique to Amazon” advantage of its Authenticated Graph, which is its core identity data set. The Authenticated Graph spans shopping and credit card data, streaming media audience and payment data and web-browsing data via the Amazon DSP, since there are still third-party cookies and other identity signals in the programmatic bidstream.
Other measurement providers and ad platforms must stitch together their measurement identity data from many sources, which means “the fidelity of that signal will be decreased,” Tong said.
What’s more, she added, Amazon’s Authenticated Graph has a deeper historical outlook. The reason Amazon can quantify the long-term value of a given new customer is that its graph isn’t just there to match IDs and secure an attributed conversion but track how those new customers shop (or don’t) with the company over time. Amazon’s longer year-over-year perspective is also what allows it to extend its measurement lookback window.
Clearly, Amazon benefits from the trend toward more outcome-based buying by TV advertisers. Although buzzwords can be tricky and capricious.
For example, when asked whether Amazon has any measurement news related to the upfronts and marketing mix modeling, Tong noted that the MMM product falls under a different measurement director’s purview. Likewise, incrementality testing, another measurement buzzword this year, “has an entire team that focuses” on just that marketing application, she said.
Which affords some hope for the enduring value of third-party attribution vendors and specialists despite Amazon’s myriad advantages in measurement with its Authenticated Graph. There are enough outcomes out there for everyone.
