What Retail Media Advertising Actually Means (And Where Amazon Ads Fits)

What retail media actually is

Retail media is advertising bought inside a retailer's own digital storefront — search results, product pages, category pages, sometimes the app — priced against that retailer's first-party purchase data rather than a third-party cookie. It sits inside a live shopping session, next to an actual decision to buy, which is why it behaves differently from a banner ad next to a news article.

Amazon Ads is the largest, most mature version of this, but retail media is a category now, not one company. Walmart Connect, Target Roundel and Kroger Precision Marketing run the same basic model on their own traffic. Amazon has also started licensing its ad technology to other retailers outright — that's Amazon Retail Ad Service, currently running on iHerb, Weee!, Oriental Trading and a small group of others, on their sites, not on Amazon.com.

Where Amazon Ads fits inside retail media

Inside Amazon's own store, "Amazon Ads" is three layers stacked on top of each other, and each layer answers a different question:

• Sponsored Products and Sponsored Brands answer "which product or brand wins this specific search," auctioned per keyword, paid per click, reported on last-click sales.
• Sponsored Display extends that into retargeting — showing an ad to someone who viewed a product but didn't buy — on and sometimes off Amazon.
• Amazon DSP is a different animal: programmatic buying of display, video and audio against audience segments, running on Amazon-owned properties (Fire TV, IMDb, Twitch) and the open web and app exchanges Amazon has inventory access to. It's bought on a CPM, not auctioned per keyword, and it's built for reach and audience building, not for winning a single search result. It can be bought self-service, through Amazon's managed service, or via an agency seat — what a self-service seat requires, and who actually qualifies is its own question.

Full Circle, the management company behind this page, manages Amazon programs for brands as different as HexClad, Ridge, BK Beauty, Beardbrand, Epic Gardening and The Woobles — cookware, wallets, beauty, grooming, gardening and DTC crafts. The layers above stay the same regardless of category; what changes is which layer earns the budget first.

A worked example: turning a media budget into a number that means something

Retail media reporting is full of terms — CPM, CPC, ROAS, attributed sales — that sound precise and say nothing until you see them stacked end to end. Here's one real stack: across 30 advertisers running Amazon DSP in July 2026, the media bought 78.4 million impressions at a $4.00 CPM. Those impressions produced a blended $1.42 cost-per-click. Clicks and views together converted into 57,137 attributed purchases at a blended $5.49 cost-per-acquisition, 20.1% of them from shoppers who had never bought the brand before. The overall return across that book was 6.04x ROAS.

Read it end to end and each figure does a specific job. CPM tells you what the impressions cost. CPC tells you whether the creative and targeting earned attention. CPA tells you whether that attention turned into a sale at a price you can live with. The new-to-brand share tells you whether the spend grew the customer file or just re-sold people who already buy from you. ROAS is the number everyone quotes first, and it's the one that tells you the least on its own — it's an average of a funnel, not a description of one.

The mistake almost every retail media report makes

The biggest one: treating last-click attribution as proof an ad worked. Last-click hands 100% of the credit for a sale to whichever ad a shopper touched right before buying — even if they were already going to buy, and even if a display impression they saw three days earlier is what actually moved them. Click order isn't causation. It never was.

The related mistake, one we've made ourselves before fixing the reporting: running Sponsored Products and Amazon DSP on the same account and letting both platforms claim the same purchase in separate dashboards. Add the two ROAS numbers together and the business looks like it's printing free money. It isn't — it's one sale counted twice. The only way to catch this reliably is to reconcile both channels in Amazon Marketing Cloud, where a purchase can be attributed once across the whole account, not once per platform. That's the only honest way to say whether the display spend actually added anything on top of what sponsored ads were already going to sell.

When the numbers don't look right, what to check first

If a retail media report comes back looking wrong, check these before deciding the campaign failed:

• Is the attribution window actually consistent between the two periods you're comparing? A 14-day window this month against a 7-day window last month will move ROAS on its own, with no change in performance.
• Are Sponsored ads and DSP reconciled, or just stacked side by side? Two strong-looking dashboards can still describe one incremental sale.
• Was there a holdout or matched control, or is the "lift" just the gap between a promoted period and a normal one? Without a control group, a seasonal spike and an ad-driven spike look identical on a chart.
• Did the setting that's supposed to fix this already exist and simply wasn't turned on — new-to-brand reporting, cross-channel dedupe, frequency caps? A large share of "the number is wrong" turns out to be "the report was never configured to answer this question."

If none of that explains it, the campaign probably did underperform. The honest next move is a rollback, not a bigger budget hoping it averages out over time.

Where Dr. DSP fits

Dr. DSP runs Amazon DSP — Demand-Side Platform, Amazon's programmatic buy for display, video and audio, not the Delivery Service Partner courier franchise — as a managed product for Fable 5, from Full Circle. Full Circle has managed more than $500 million in Amazon ad spend across 100+ brands, and There are 70+ brands live across the Full Circle and reMKTR group right now. None of that replaces the checks above — it's what reconciling DSP and sponsored ads in Amazon Marketing Cloud, testing with holdouts, and attaching a rollback trigger to every change looks like when it's someone's full-time job. There's no published price: a demo, the first 30 days free, and pricing set on a call against your real budget and scope, with Orbit, the reporting suite, included either way. Whether or not you ever call, the point of this page is that the difference between an ROAS number and a proven incremental one is the whole game — and most retail media dashboards never show you that gap.