Amazon Marketing Cloud is free and already in your ads console. Here is what it is actually for.

What Amazon Marketing Cloud is, in plain terms

Amazon Marketing Cloud — AMC, and everyone says AMC — is a secure analytics environment Amazon describes as a clean room, built on AWS Clean Rooms. You get access to what Amazon's own product page calls hundreds of fields about your Amazon Ads campaigns, including ad-attributed impressions, clicks and conversions, and you can upload your own pseudonymous data and join it to Amazon's signals. Amazon says it is available at no cost to eligible advertisers.

The word that matters is clean room. AMC takes only pseudonymised inputs and gives back only aggregated, anonymous outputs, with what Amazon calls built-in aggregation thresholds. You will never see a shopper. You will never export a list of names. What you get instead is the ability to ask a question that spans individual shoppers and have the answer come back as a number.

That sounds like a technicality and it is the entire point. Every report you have today — the search term report, the bulk file, the campaign dashboard, Search Query Performance — is aggregated before you see it, by campaign, by keyword, by ASIN, by day. None of them can tell you that the person who bought on Thursday had already clicked a Sponsored Products ad on Monday and scrolled past a Sponsored Display ad on Tuesday. AMC can, because it joins those events underneath the aggregation rather than on top of it.

Why this suddenly applies to sellers who have never touched Amazon DSP

For most of its life AMC was a thing other people had. Access was tied to Amazon DSP, which meant display and streaming budgets, an account executive, and a minimum that ruled out anyone running a normal Sponsored Products account. That changed in two steps, and the second step is the one worth knowing about.

The first step was on 15 October 2024, when Amazon expanded eligibility so that all Amazon DSP direct advertisers and all partners registered with the Amazon Ads Partner Network could register for AMC — including partners serving clients who only use sponsored ads. In practice that still meant going through somebody else.

The second step was 16 September 2025. Amazon's launch note says sponsored ads advertisers can now access AMC directly within the ads console, in the left-hand navigation bar under Measurement & Reporting, and explicitly says this works without requiring partner or representative intervention. Amazon's own description of AMC for sponsored ads puts the eligibility plainly: it is available to advertisers running Sponsored Products, Sponsored Display, Sponsored Brands or Sponsored TV campaigns, and setup takes a few clicks. The marketplace list in the launch note is long — the US, Canada, Mexico, Brazil, the UK, Germany, Spain, Italy, France and a dozen more European markets, several Middle Eastern markets, Australia, New Zealand, Japan, India and Singapore among them.

So if you are running Sponsored Products on amazon.com today, the honest answer to "am I big enough for Amazon Marketing Cloud" is that eligibility is no longer the question. Open the console and look under Measurement & Reporting. Either it is there or it is not, and that takes thirty seconds to establish.

The question your normal reports cannot answer

Here is the practical gap AMC fills. Your campaign report says Sponsored Brands returned a 4.1 ACOS-equivalent and Sponsored Display returned something worse, so you move budget from Display to Brands. Six weeks later total sales are flat and you cannot work out why. The reason this happens so often is that attribution in the normal reports assigns each sale to one campaign, and shoppers do not shop in one campaign.

Amazon has been putting the raw material for that analysis into AMC for years. Sponsored Brands engagement and conversion signals landed there on 13 January 2023, and Amazon's description of what they unlock is exactly the list of things the standard reports will not do: measure total and incremental reach across your sponsored ads portfolio, analyse sales from audiences who saw or engaged with Sponsored Brands ads, find high-performing keywords across both Sponsored Brands and Sponsored Products, and map how media exposure contributed to a conversion.

The first thing most accounts should ask, then, is an overlap question: of the shoppers who bought, how many were reached by one ad type only, and how many by two or more? You do not need a sophisticated model to use the answer. If three-quarters of your converters only ever saw Sponsored Products, the campaign type you were about to cut is not doing the job you hoped, and the budget argument is settled. If the opposite is true — converters who saw two ad types convert at a meaningfully higher rate than those who saw one — then the campaign you were measuring in isolation is propping up the one that gets the credit.

Treat that as directional, not causal. Shoppers who see more of your ads are often shoppers who were already searching more, and an overlap figure does not separate the two. It is a good enough reason to stop cutting a campaign on single-campaign ACOS alone; it is not proof the second ad caused the sale. If you want proof, you need a holdout, and that is a different and slower exercise.

You no longer have to write SQL to get the first answer

The old objection to AMC was that it was a SQL console, and most people running their own account do not write SQL. That objection has weakened considerably. Amazon's description of the sponsored ads experience is that the AMC homepage offers no-code analysis templates that return visualised insights such as campaign performance and audience insights, with a low-code use case library behind them and the full query editor behind that.

On top of that, on 31 March 2026 Amazon added AI assistance inside the query editor itself: AI-powered controls on a right-click context menu, the ability to send selected SQL to the Ads Agent chat to refine it conversationally, and a side-by-side diff view of what the AI changed. Amazon says this is available to anyone with access to AMC.

This changes the realistic first session from "learn SQL" to "run a template and read the chart". It does not change the thinking, which is still the hard part. A query that returns a number you had no plan for is a waste of an afternoon. Decide what you would do differently under each possible answer before you run anything — if both answers lead to the same action, the query is not worth running.

Twenty-five months of history changes which questions are worth asking

On 11 November 2025 Amazon expanded the AMC ad traffic lookback window from 13 months to 25 months. It launched as an open beta in the US and Canada, with Amazon saying expansion to the other listed locales — the UK, Germany, Italy, France, Austria, Turkey and Japan — would happen in Q1 2026, and that API users need no endpoint changes to get it.

Thirteen months sounds like a year of history and is not. It means that on any given day, last year's equivalent week is either just inside the window or just outside it, and a clean year-over-year comparison of a seasonal period was never reliably available. Twenty-five months means you can put Q4 2025 next to Q4 2024 and look at both in full.

For a seasonal catalogue that is the single most valuable thing in this whole list, and the timing is useful: you are deciding Q4 budgets now with last Q4 as your only real evidence. Being able to ask whether the pattern you remember from last year was actually a pattern, or a one-off, is worth more than another bid adjustment.

Turning an answer into an actual campaign change

Analysis that does not end in a change to a campaign is a hobby. The bridge in AMC is audiences, and this is the part most sellers do not know exists. On 15 October 2024 Amazon extended AMC audiences to sponsored ads: advertisers can use rule-based or lookalike audiences built in AMC for Sponsored Display targeting, and as bid boosts in Sponsored Products and Sponsored Brands.

Read that second clause again, because it is the one that applies to nearly everyone. A bid boost means you keep the campaign you already run and simply bid more when the shopper in front of you belongs to a group you defined — people who viewed the product and did not buy within a fortnight, people who bought a complementary ASIN, people who bought twelve months ago and are due to repurchase. You are not building a new campaign structure. You are putting a multiplier on the one you have, aimed at a segment the keyword alone cannot identify.

A caution on size. Amazon's launch notes for AMC audiences do not publish a minimum audience size, and third-party figures circulate that we are not going to repeat here. What is safe to say is that AMC applies aggregation thresholds by design, so an audience defined too tightly can come back empty or fail to build, and you will find that out after you have built it rather than before. Size the segment with a counting query first, then build it. Start with a definition that is obviously large — all product-page viewers in 90 days who did not purchase — and narrow it once you have seen it work.

What to do on Monday

Open the ads console and look in the left-hand navigation under Measurement & Reporting. If Amazon Marketing Cloud is there, complete the setup; it costs nothing and the instance starts collecting against your history rather than from today. If it is not there, check which marketplace and account you are signed into before concluding anything, because eligibility is per marketplace.

Then run exactly one thing: an overlap analysis across your ad types for the last 90 days, from the no-code templates. Write down, before you look, what you expect the answer to be and which campaign you would cut if you are right. Most people find the result disagrees with them, and the disagreement is the value.

Third, if your catalogue is seasonal, use the longer lookback for one year-over-year question about last Q4 rather than browsing. One question, written down, answered, acted on.

And a caution worth keeping. AMC will not tell you whether an ad caused a sale, it will not give you anything at the individual shopper level, and it will not fix an account whose structure or listings are the real problem. It is a measurement tool that makes cross-campaign questions answerable for the first time without a DSP budget. If your Sponsored Products search terms are still full of waste, fix that first — the cheaper insight is always the one you already have the data for.