Amazon Marketing Cloud in 2026: What It Is, When It Pays, and How Growing Brands Actually Use It

Your Amazon Ads console shows what each campaign did last week. It can't show you that 70% of your conversions started with a Sponsored Brands ad you nearly cut for a high ACOS, got reinforced by a DSP retargeting impression a few days later, and only closed on Sponsored Products. The console sees the last click. The customer saw the whole sequence. That blind spot is exactly what Amazon Marketing Cloud was built to remove.

AMC isn't new. What's new is that 2026 is the year it stopped being a tool reserved for the biggest advertisers. Amazon made its first-party signals free through the end of the year, added a generative AI layer that writes the queries for you, and stretched the purchase lookback out to five years. For brands doing real volume, the effort finally pays. Below is what AMC actually is, what it costs, when it's worth your time, and the specific questions we run it to answer across our portfolio.

What Amazon Marketing Cloud Actually Is

AMC is a privacy-safe clean room. You feed it pseudonymized data, it returns aggregated and anonymous results, and whatever you upload stays inside your own dedicated instance. Amazon can't access it, and you can't export individual user records out of it. Nobody is handing raw customer lists back and forth. You ask questions of the combined data and get answers, not files.

The reason that matters: AMC stitches Sponsored Products, Sponsored Brands, Sponsored Display, and DSP into one dataset keyed to the same shopper. Then it lets you query the path between them. You get 25 months of ad-signal history to work with. A newer dataset called Amazon Retail Purchases extends the store purchase signal from 13 months to five years, which changes the math for durable goods, replenishables, and anything seasonal where the repeat cycle is long.

Two 2026 additions lower the barrier further. AMC now has a natural-language layer: describe the audience or the question in plain English and it writes the SQL for you, cutting query development from hours to minutes. And Custom Models for Audiences, built on AWS Clean Rooms and now generally available, let you train your own models on first-party data combined with Amazon signals without moving sensitive data out of your AWS environment. Audiences you build in AMC activate directly across Sponsored Ads, DSP display, and video. You're not just measuring. You're turning the finding into a media buy.

What AMC Costs, and What It Actually Costs

Access is free. If you run Sponsored Ads, you can self-serve into AMC. If you run Amazon DSP, you request access through your Ad Tech account executive. Through December 31, 2026, Amazon's first-party paid-feature signals are also free, which previously sat behind a subscription.

The paid features matter because one of them, Amazon Flexible Shopping Insights, is the only way to see organic sales, organic adds-to-cart, and organic detail-page views inside AMC. That's the data that tells you whether your ads are driving net-new demand or riding on organic that was already there. Third-party signals, like Experian segments, still require a paid subscription, and paid-feature pricing is customized per account with no published rate card. So don't budget around a sticker price you found in a blog post. There isn't one.

Here's the honest part most vendors skip. The real cost of AMC isn't access. It's analyst time. Someone has to write the queries, read the outputs, and then actually change a budget based on what they find. For a mid-size brand that's roughly 8 to 16 hours a month of skilled work. A clean room full of insights nobody acts on is a cost, not an asset. We've audited brands paying for the analytical horsepower and doing nothing with it. That's the most expensive way to use AMC.

When AMC Pays Off

AMC rewards scale and complexity. The more campaign types, SKUs, and channels you run, the more there is to untangle, and the more a path-level view is worth. A single-SKU brand running one Sponsored Products campaign doesn't have a path problem worth a clean room. A multi-SKU brand running full-funnel across Sponsored and DSP does.

Monthly Amazon ad spend	Is AMC worth it?	What you'd actually learn
Under $50K	Not yet	Fix campaign structure and listings first. The console has what you need.
$50K to $200K, single-SKU	Usually wait	Path complexity is low. Limited payback on analyst time.
$200K to $400K, multi-SKU or multi-channel	Yes	Path-to-purchase, audience overlap, new-to-brand by segment.
$400K and up	Clear yes	Incrementality testing, custom audiences, full-funnel budget modeling.

These are rules of thumb, not hard lines. A brand spending $150K a month across 40 SKUs on Sponsored and DSP can get more from AMC than a brand spending $300K a month on one hero product. Spend is a proxy for complexity. Complexity is what AMC actually charges against.

The AMC Queries Worth Running

AMC ships with instructional query templates, and plenty of agencies publish copy-paste SQL libraries. Most of the value, in our experience, comes from four questions. Not forty. Four.

Query	What it answers	The decision it drives
Path-to-purchase	Which ad types start, assist, and close conversions, and in what order	Stop cutting upper-funnel campaigns by last-click ACOS alone
Audience overlap	How much of one campaign's audience is also served by another	Cap redundant frequency and move budget to incremental reach
New-to-brand by segment	Which audiences deliver first-time buyers versus repeat buyers	Tilt acquisition spend toward segments that grow the customer base
Incrementality (exposed vs unexposed)	Whether ad-exposed shoppers buy more than comparable unexposed ones	Separate true incremental sales from purchases you'd have won anyway

Two of these tend to change budgets immediately. Path-to-purchase routinely shows that a campaign type teams want to cut, often Sponsored Brands or DSP, is initiating a large share of the journeys that close elsewhere. We've seen accounts where roughly 70% of conversions began with a Sponsored Brands touch the console credited entirely to Sponsored Products. Cut the SB campaign on its standalone ACOS and you'd quietly kneecap the SP campaign that looks like your winner.

Audience overlap is the other one. When 40% of a DSP audience is also getting hit by Sponsored Display, you're paying twice to reach the same person and pushing frequency past the point of return. The overlap report makes that waste visible, and the fix is usually a fast budget reallocation toward audiences you're not already saturating. Neither finding is exotic. Both are invisible without AMC.

AMC Solves the Amazon Slice, Not the Whole Problem

It's worth being clear about a limit. AMC is an Amazon clean room. It sees Amazon and DSP signals. It does not see your Meta spend, your Google campaigns, or your Shopify checkout. For a brand running five to seven ad networks, AMC answers the Amazon question deeply and leaves the cross-channel question open.

That open question is the one keeping marketing directors up at night. In recent retail-media research, 75% of brands named incrementality as their single biggest measurement challenge, 59% named cross-channel attribution, and only 15% said they felt very effective at measuring retail-media performance at all. Each platform measures differently. Amazon reports by SKU, Meta by audience, Google by keyword, and the three views rarely reconcile. AMC is a precise answer to one slice of that. The full answer pairs AMC with portfolio-level methods like geo-based holdout tests and media mix modeling, so the Amazon read sits inside a cross-channel one instead of competing with it.

How We Use AMC Across the Portfolio

We treat AMC outputs as decisions, not dashboards. For accounts at the right scale, path-to-purchase and overlap reports run on a regular cadence, and we reserve incrementality testing for the largest line items, the ones where being wrong costs the most. Across more than $450M in managed Amazon revenue, 100+ brands, and $7M+ in annual ad spend, the pattern holds: the brands that win with AMC aren't running the most queries. They're acting on the three or four that change where money goes.

The connection to results is direct. When path-to-purchase data stops a team from starving the campaigns that initiate demand, upper-funnel survives, the closers keep converting, and total advertising cost of sale trends toward the 12% range we target rather than spiking every time someone optimizes a single campaign in isolation. That's how a portfolio averages an 89% sales lift. Not from one clever query, but from consistently funding what's incremental and defunding what isn't. If you want help standing this up, our Amazon advertising team and full-service management group do exactly this work.

The Bottom Line

AMC is free to access, genuinely powerful, and useless if it just generates reports. The brands getting an edge from it in 2026 are the ones spending enough across enough campaign types that the path actually matters, and disciplined enough to move budget when the data says to. If that's you, and you'd rather have an operator run it than build the SQL muscle in-house, tell us what you're spending and we'll tell you whether AMC is worth it for your account.