Amazon opens MMM API publicly

Marketing mix modeling is one of those boring-sounding tools that quietly decides where big ad budgets go. It tries to answer a hard question — which channels actually drove sales, including the ones that don’t get a neat click or conversion tag. Amazon’s news is that this is now easier to pull into real workflows. In May 2026, Amazon Ads moved its Marketing Mix Modeling API from beta to general availability, opening programmatic access to its MMM data feed across 14 countries. (advertising.amazon.com) ### What is Amazon actually opening? Amazon is opening an API for MMM data, not a brand-new measurement product from scratch. The feed gives advertisers and partners aggregated Amazon sales and media-performance signals that they can request asynchronously, then download as report files. Amazon says those files are the same outputs that come from the MMM application inside the Amazon Ads console. (advertising.amazon.com) ### Why does “API” matter here? Because MMM usually breaks at the handoff. A marketing team might have a model, a consultant, and a spreadsheet process, but getting clean weekly or daily inputs into that system is often manual and slow. API access means brands, agencies, and software vendors can pull the data directly into t(advertising.amazon.com)ns MMM from an occasional project into something closer to operational reporting. (advertising.amazon.com) ### What data is in the feed? The important part is the mix of advertising and retail signals. Amazon frames the feed as a way to give advertisers the sales and media inputs they need to estimate channel impact on Amazon outcomes, including omnichannel learnings. That matters because MMM works best when it can see both s(advertising.amazon.com)re this with daily-grain metrics, which makes the data easier to line up with other publishers and faster-moving models. (advertising.amazon.com) ### Which markets are included? The rollout covers 14 countries: the US, Canada, Italy, Spain, France, Germany, the UK, Mexico, Brazil, Japan, Australia, the UAE, India, and Saudi Arabia. That list is a bigger deal than it sounds. MMM gets much more useful when multinational brands can use one access method across regions instead of running separate local workarounds. (advertising.amazon.com) ### Who is this really for? Not every seller on Amazon. This is mainly for larger advertisers and the partners around them — agencies, measurement vendors, and in-house data teams that already run MMM or want to. Amazon’s own documentation talks about advertisers and their partners using the feed to derive marketing insights, which is a clue that the tar(advertising.amazon.com)ampaign budget in the UI. (advertising.amazon.com) ### Why wasn’t the console enough? A console is fine for looking things up. It is bad for building repeatable systems. If a brand wants to refresh models every week, compare markets, join Amazon data to TV, paid social, search, pricing, and promotions, and then feed the result into planning software, exports become the bottleneck. General availability al(advertising.amazon.com)ion use, not just testing. (advertising.amazon.com) ### How does this fit with Amazon’s bigger measurement push? Basically, Amazon has been building out a stack. The core Ads API handles campaign management and reporting. Amazon Marketing Cloud gives more advanced analytics and clean-room style workflows. The MMM API sits in a different slot — it is about feeding top-down eco(advertising.amazon.com)ent teams. (advertising.amazon.com) ### Bottom line? The catch is that an API does not solve MMM by itself — brands still need models, analysts, and good non-Amazon data. But Amazon just removed one of the most annoying parts: getting consistent Amazon inputs into those models at scale. For retail marketers trying to prove what actually moved sales across channels and countries, that is a real upgrade. (advertis([advertising.amazon.com)elease-notes))