Amazon Marketing Cloud: What It Is and the Queries Worth Running.

Amazon Marketing Cloud (AMC) is a privacy-safe clean room, hosted on AWS, where an advertiser writes SQL against event-level data about its own Amazon ad impressions, clicks and conversions, and gets back aggregated results only. It answers questions the standard Ads console reports cannot, such as which combinations of ad types lead to a purchase and how long that takes. It is free to use for eligible advertisers, but it is only as useful as the questions you ask and the decisions you make with the answers.

We query AMC for our own Amazon brands, so this guide is written from the query editor. It covers what AMC is and is not, access requirements, the queries we run first, AMC audiences, and a plain test for whether a brand is ready.

What Amazon Marketing Cloud Actually Is.

A clean room is a controlled environment where two parties' data can be analyzed together without either side seeing the other's raw records. In AMC, Amazon provides pseudonymized, event-level signals for your campaigns: impressions and clicks from Sponsored Products, Sponsored Brands, Sponsored Display and Amazon DSP, plus conversion events such as detail page views, add to carts and purchases. Each AMC instance only contains data for the advertiser accounts connected to it.

You never see a row about one shopper. You write a query, AMC runs it, and the output is aggregated. Any row that represents too few distinct users is dropped or suppressed before you see it, and there are further protections against queries designed to single people out. This is the single most important thing to understand: AMC answers questions about groups, and small groups disappear from the output.

Access Requirements.

AMC began as a tool for Amazon DSP advertisers, usually reached through an agency or Amazon's managed service. Amazon has since broadened access so that many Sponsored Ads advertisers can open an AMC instance from the Amazon Ads console. Eligibility, regions and the exact sign-up path change over time, so check the current requirements in the console rather than trusting an old checklist.

In practice you need three things beyond eligibility: advertiser accounts connected to the instance, someone who can write and debug SQL, and enough ad volume that your questions return rows. AMC itself does not charge a fee to run queries, though Amazon sells optional paid insight add-ons, and your time is the real cost. If you are weighing whether to staff this in house, our guide to what a real Amazon marketing agency engagement looks like covers the skills involved.

The Queries Worth Running First.

AMC ships with an instructional query library: templates written by Amazon that you copy, adjust dates and filters on, and run. Start there. Writing from scratch is for the second month. These are the questions we answer first on any account.

Path to purchase groups converting users by the combination of ad types they were exposed to and reports conversion rate for each combination. It is the query that tells you whether shoppers who saw a Sponsored Brands video and then a Sponsored Products ad buy at a higher rate than those who saw Sponsored Products alone. Time to conversion measures the gap between first exposure and purchase, which is how you set a sensible remarketing lookback instead of guessing. New-to-brand in AMC can be cut in ways the console cannot, for example by combinations of campaigns. Overlap between Sponsored Ads and DSP shows how much of your DSP audience also saw your search ads, which is the core question behind any DSP budget.

Running Your First Query, Step by Step.

AMC Audiences.

AMC is not only for analysis. You can define an audience with a query, for example shoppers who added a product to cart but did not purchase within a window, or buyers of one product who have not bought a complementary one, and send it to Amazon DSP. Amazon has also extended AMC audiences to Sponsored Ads in some forms, such as audience targeting in Sponsored Display and audience bid adjustments. Check the console for what is supported on your account, because this area changes.

Audience queries follow the same privacy rules: the audience itself must be large enough to be activated, and you never see who is in it. The practical limit is the same as for analysis. A small brand's cart abandoners in a short window may be too few to use.

What Usually Goes Wrong.

• Slicing too finely on the first run and concluding AMC has no data, when the rows were suppressed by the aggregation threshold.
• Opening AMC for an account that runs only Sponsored Products at modest spend, where the console reports already answer the useful questions.
• Treating path to purchase as causal proof. Shoppers who see more ad types are often already more interested, so compare against holdouts or at least read it with that bias in mind.
• Running interesting queries once and never scheduling them, so there is no trend and no comparison over time.
• Mixing data from different advertiser accounts or date ranges without noting it, and then comparing results that were never like for like.
• Building audiences that are too narrow to activate, then waiting on campaigns that never serve.

When a Brand Is Ready for AMC.

A brand is ready for AMC when three things are true. It runs at least two or three Amazon ad types, usually Sponsored Products plus Sponsored Brands or Sponsored Display, and ideally DSP. It spends steadily enough that weekly groups by ad type hold real volume. And it has a question the console cannot answer, such as whether upper-funnel spend is producing new customers or just taking credit for sales Sponsored Products would have made.

If none of that is true, the better use of the same hours is usually the basics: clean Sponsored Products structure, strong listings, and Brand Analytics. AMC rewards maturity. It does not create it.

Where the Clean Room Stops and the Work Begins.

AMC gives you a query editor and a library. The work is deciding which questions matter, writing SQL that respects the privacy rules and returns usable rows, scheduling queries so trends are visible, turning the answers into budget moves across Sponsored Ads and DSP, and building audiences large enough to activate. It is analyst work tied to media buying, and the value only appears when the same people do both. That combined analysis and media work is what we do through Amazon DSP and Marketing Cloud management.