How Amazon Marketing Turns Into Amazon Sales

What "amazon marketing and sales" actually means

On Amazon, marketing and sales aren't two departments handing off a lead. They're two stages of the same log file. Marketing is anything that puts a listing in front of a shopper: sponsored search ads, Sponsored Brands and Sponsored Display, Amazon DSP for display, video and streaming TV, plus the unpaid layer — titles, images, A+ content, and backend keywords that decide whether the listing shows up at all. Sales, in Amazon's own reporting, almost never means total revenue on the product. It means an attributed purchase — one that Amazon's system connects back to an ad exposure inside a defined window.

That distinction matters because it's where most confusion starts. A brand can grow category sales while its attributed sales look flat, or the reverse. The gap is the shoppers who bought without ever touching an ad — direct search, repeat buyers, word of mouth — versus the ones the system can trace back to a click or a view.

Sponsored ads and Amazon DSP aren't ranked versions of the same tool. Sponsored ads sit inside Amazon's search results and product pages, aimed at people already looking. DSP runs display, video and OTT inventory on and off Amazon, aimed at people who haven't searched yet. Treating DSP as "sponsored ads but bigger" is a common misread — it's a different stage of the funnel, doing a different job.

How marketing spend becomes an attributed sale — a worked example

Here's how the numbers actually chain together, using a real book of business: across 30 advertisers reMKTR managed in July 2026, the campaigns ran 78.4 million impressions at a $4.00 CPM. Multiply impressions by CPM and you get spend on the display side of the ledger — roughly $313,600. That same spend, spread across 57,137 attributed purchases, works out to a blended cost per acquisition of $5.49. The math holds together because it's the same book measured two ways: cost per thousand impressions on one side, cost per purchase on the other.

The blended cost-per-click across that book was $1.42 — worth flagging because the $0.41 CPC figure that circulates in this category is an online-video-only number, not a whole-book average. If a vendor or a listicle quotes a CPC without saying which channel it came from, assume it's the cheapest slice, not the average.

Return on ad spend for that book came to 6.04x, measured across all 30 advertisers rather than pulled from the best-performing line item. One in five of those attributed purchases — 20.1% — came from a shopper new to the brand, which is the number that tells you whether the spend grew the customer base or just harvested people who were going to buy anyway.

Sponsored ads vs DSP: two jobs, one sales number

Sponsored Products captures demand that already exists — someone typed a search term close to your product. Sponsored Brands and Sponsored Display sit slightly earlier, catching comparison shoppers. DSP sits earliest: someone hasn't searched for the category yet, but the audience signal says they're likely to.

All three report into the same sales metric, which is exactly the problem. If a shopper sees a DSP video ad on Monday and clicks a sponsored ad on Wednesday, both campaigns can claim the purchase in their own dashboard. Add the two ROAS numbers together and the business looks like it's converting more spend into sales than it actually is.

This isn't a hypothetical failure mode — it's the default behavior of last-click reporting, and it's cost real brands real budget decisions. The only way to see it clearly is to reconcile DSP and sponsored ads in the same measurement layer, such as Amazon Marketing Cloud, so overlapping exposures get counted once instead of twice.

The mistake that inflates results: double-counting across ad types

We've made this mistake too. Early on, moving budget into a DSP retargeting pool made the blended ROAS look better within a quarter — but a chunk of those "conversions" were shoppers who'd already added the item to cart before the ad ever served. The dashboard said the campaign worked. A holdout comparison against a matched group that saw no ad said it added almost nothing.

Last-click attribution can't tell those two situations apart. It rewards whichever campaign happens to sit closest to the checkout, regardless of whether that campaign caused the checkout. Holdout tests and matched controls can tell the difference, because they compare exposed shoppers against otherwise-identical unexposed ones and measure the gap.

The practical rule: if a reported ROAS number can't survive being checked against a holdout or a matched-market test, treat it as a directional signal, not a fact you'd defend to a CFO.

When the sales number looks wrong — what to check first

If attributed sales drop and nothing about the business changed, work through this order before touching the budget:

• Attribution window overlap — check whether Sponsored Products and DSP are both claiming the same purchase inside their respective windows.
• ACOS vs TACOS confusion — a rising ACOS on one campaign can coexist with a falling total advertising cost of sales across the account; look at both before reacting.
• New-to-brand rate — a campaign with strong ROAS but a low new-to-brand percentage is likely retargeting people who were already going to buy, not growing anything.
• Seasonality and stockouts — a dip that lines up with an inventory gap isn't a marketing problem, and no amount of bid adjustment will fix it.

If you've checked all four and the number is still bad, the fix isn't a bigger budget or a new bid strategy. It's a test: hold out a matched segment, run the campaign against the rest, and compare. If the holdout group converts at nearly the same rate as the exposed group, the spend isn't earning its keep — no dashboard metric will tell you that as reliably as the test will.