Amazon this month moved its Marketing Mix Modeling API to general availability, ending a beta period that had restricted access to the data feed underpinning the company's measurement infrastructure. The announcement, published in the May 2026 section of the Amazon Ads Advanced Tools Center release notes, marks the first time that all advertisers and their partners can access Amazon's aggregated advertising and retail signals programmatically for use in their own MMM frameworks - without navigating beta constraints or waiting for invitation-based access.

The scope is notable. According to the release notes, the MMM API now enables programmatic access to Amazon's MMM data feed across 14 countries: the United States, Canada, Italy, Spain, France, Germany, the United Kingdom, Mexico, Brazil, Japan, Australia, the United Arab Emirates, India, and Saudi Arabia. That geographic footprint spans four continents and covers most of the major markets in which Amazon operates a meaningful retail and advertising business. Whether an agency is running a model for a consumer packaged goods brand in Germany or a direct-to-consumer retailer in Japan, the same API endpoints now apply.

What the MMM API actually does

At its core, the MMM API is an asynchronous API. It does not return data immediately upon request. Instead, advertisers submit report requests, and the system processes them in the background before making the output available. According to the technical documentation in the Amazon Ads Advanced Tools Center, the responses contain the same report files that would be generated by the MMM application inside the Amazon Ads console. The API essentially replicates the console's output in a format that can be consumed programmatically, without requiring a human to log in and manually download reports.

This design has a practical implication. Teams that have been accessing MMM data through the console interface - downloading reports manually and feeding them into modeling pipelines - can now automate that extraction step entirely. For agencies managing multiple advertiser accounts, or for brands running continuous MMM updates rather than periodic batch analyses, that shift from manual to programmatic access represents a meaningful change in operational overhead.

The underlying data combines Amazon's advertising signals with retail signals. This combination is what distinguishes Amazon's MMM data from what most other platforms offer. A typical advertising platform can supply media performance metrics: impressions, clicks, spend, and attributed conversions. Amazon's retail infrastructure adds a second layer - actual sales data from its marketplace, including signals that reflect consumer purchase behavior across the platform's product catalog. For MMM practitioners, having both media and retail signals from a single source, already aligned and formatted for modeling, addresses a data assembly problem that typically requires considerable manual effort.

New operations added since beta

The general availability release is not simply a status change. According to the release notes, the MMM API now includes several new operations that were not part of the beta, giving advertisers greater control over brand group configurations and improved visibility into report status.

Six new endpoints are documented. The first two handle brand group overrides: POST /mmm/v1/brandGroupOverrides creates overrides for brand groups, while POST /mmm/v1/brandGroupOverrides/delete removes them. A third endpoint, POST /mmm/v1/brandGroupOverrides/list, retrieves the current list of overrides. Two GET endpoints address product and campaign retrieval within brand groups: GET /mmm/v1/brandGroups/{brandGroupId}/products retrieves the list of products associated with a given brand group, and GET /mmm/v1/brandGroups/{brandGroupId}/campaigns retrieves the campaigns linked to those products. The sixth new endpoint, POST /mmm/v1/reports/list, allows advertisers to list reports and check their generation status.

The brand group override endpoints are worth unpacking. In Amazon's MMM framework, a brand group is a logical container that groups related products and campaigns for modeling purposes. Overrides allow advertisers to adjust how those groupings are defined - useful when the default grouping logic does not match the way a brand structures its product portfolio internally. The ability to create, delete, and list those overrides via API, rather than only through the console, removes a friction point for advertisers whose brand architecture is complex or changes frequently. The GET /mmm/v1/brandGroups/{brandGroupId}/campaigns endpoint in particular adds transparency: it lets advertisers verify programmatically which campaigns are feeding into a specific brand group's model, without having to cross-reference through the console UI.

The report status endpoint - POST /mmm/v1/reports/list - addresses an operational challenge inherent to asynchronous APIs. Because the MMM API processes reports in the background, a caller has no immediate confirmation that a report is ready. The new listing endpoint fills that gap, letting developers build polling logic or status dashboards that track report generation without depending on manual checks.

Eligibility and geographic reach

According to the technical documentation, MMM is available through the North America region only from an API connectivity standpoint - meaning API calls must be routed through the NA region endpoint. However, the data itself is not limited to North America. The API supports requests for data from 14 individual marketplaces spread across four regions: North America (United States, Canada, Mexico), South America (Brazil), Europe (Germany, Spain, France, Italy, United Kingdom), Middle East (Saudi Arabia, United Arab Emirates), and Asia Pacific (Australia, Japan, India).

This regional API constraint with multi-market data coverage is a technical distinction that matters for implementation. Developers building integrations need to authenticate and connect through the NA region infrastructure, but can then request data for any of the 14 supported marketplaces within a single API relationship. That architecture simplifies the integration overhead compared to a model that would require separate API connections for each regional endpoint.

Why this matters for measurement teams

Marketing mix modeling has been gaining traction across the industry as privacy regulations and technical constraints have reduced the reliability of deterministic attribution. Industry research published in late 2025 found that 46.9% of marketers planned to increase MMM investment over the following 12 months, the highest planned increase among measurement methodologies. Yet adoption has remained limited: only 15% of marketing teams had actually implemented MMM, with just 8% of in-house teams possessing the required advanced analytics skills.

One persistent challenge for MMM practitioners has been data sourcing. A model is only as good as the signals fed into it, and assembling consistent, high-quality data from multiple advertising platforms - each with its own export formats, latency profiles, and definitional quirks - is itself a significant undertaking. Amazon's decision to provide a dedicated API for its MMM data feed addresses one piece of that puzzle directly. Advertisers that run substantial Amazon media investments now have a structured, programmatic path to the data they need, rather than relying on console exports or working with measurement partners who access the data through less systematic means.

The question of who builds the model remains relevant context here. A PPC Land analysis of Meta's Robyn MMM examined the structural dynamic that can emerge when a platform creates or controls the tool that guides budget allocation - a dynamic that applies equally to any measurement infrastructure built or operated by an entity with a commercial interest in the outcome. Amazon is not building the MMM models themselves here; it is providing data that advertisers and their partners feed into their own models. That is a meaningful distinction, though it does not remove the need for advertisers to think carefully about data quality, signal selection, and model validation.

The IAB published a white paper in April 2026 arguing that MMM is structurally ill-suited for retail media measurement without adaptation, noting that applying standard MMM methodologies to commerce channels can systematically misattribute sales. Amazon's data feed - which combines advertising signals with retail signals from its own marketplace - may address some of those structural limitations, since the underlying data reflects both media exposure and purchase behavior within the same ecosystem. Whether that integration resolves the attribution gaps the IAB describes, or introduces different ones, depends on how individual models are specified and validated.

Context within Amazon's broader API expansion

The MMM API general availability is one part of a pattern of programmatic infrastructure expansion that Amazon has been executing across its advertising platform. Amazon opened its advertising APIs to AI agents through the Model Context Protocol in February 2026, enabling natural language interactions with Amazon Ads API functionality. The Brand Stores management API moved from beta to general availability in February 2026, giving advertisers programmatic control over storefront content. Amazon launched the Inventory Management Unified APIs in open beta in March 2026, consolidating access to deals, supplier proposals, and inventory groups. And Amazon published a detailed technical walkthrough of its Ads Data Manager in May 2026, explaining the four-step API workflow for first-party data ingestion.

The MMM API fits within this broader pattern as a measurement-layer addition. Where the previous API expansions addressed campaign management, inventory access, and audience data workflows, the MMM API specifically targets the measurement and modeling layer - the part of the advertising stack where advertisers assess the overall effectiveness of their investments rather than executing individual campaigns. Its general availability arrival in May 2026 suggests Amazon is now extending programmatic coverage from execution infrastructure into measurement infrastructure.

For measurement teams, the availability of a stable, production-grade API - rather than a beta with uncertain timelines and potential breaking changes - changes the calculus on whether to build automated ingestion pipelines around Amazon's MMM data. Beta APIs carry operational risk; general availability carries an implicit commitment to stability that makes long-term infrastructure investment more defensible. Teams that have been waiting for that stability signal before committing engineering resources to Amazon MMM data integration now have it.

Timeline

Summary

Who: Amazon Ads, addressing advertisers, agencies, and measurement partners who use marketing mix modeling to assess advertising effectiveness across Amazon's platform.

What: The Marketing Mix Modeling API reached general availability in May 2026, providing programmatic access to Amazon's aggregated advertising and retail signals needed for MMM implementations. Six new API endpoints were added at launch covering brand group override management (POST /mmm/v1/brandGroupOverrides, POST /mmm/v1/brandGroupOverrides/delete, POST /mmm/v1/brandGroupOverrides/list), product and campaign retrieval within brand groups (GET /mmm/v1/brandGroups/{brandGroupId}/products, GET /mmm/v1/brandGroups/{brandGroupId}/campaigns), and report status listing (POST /mmm/v1/reports/list).

When: The general availability announcement was published in the May 2026 section of Amazon's Ads Advanced Tools Center release notes, documented as of May 3, 2026.

Where: The API is accessed through Amazon's North America region endpoint but covers data from 14 marketplaces: United States, Canada, Mexico, Brazil, Germany, Spain, France, Italy, United Kingdom, Saudi Arabia, United Arab Emirates, Australia, Japan, and India.

Why: Programmatic access to Amazon's combined advertising and retail signals removes the need for manual console-based report extraction, enabling automated MMM data pipelines for advertisers and partners. The general availability designation provides the stability commitment - compared to beta status - that justifies long-term engineering investment in automated integrations. The release comes amid rising industry demand for MMM as deterministic attribution methods face increasing constraints from privacy regulation and platform tracking restrictions.

Share this article
The link has been copied!