Retail Media Statistics 2026: The Numbers That Matter

Retail media statistics tell the clearest growth story in advertising: it is the fastest-growing major channel, compounding at roughly 20% or more a year per eMarketer and GroupM forecasts, with Amazon Sponsored Products CPCs clustering near $0.90–1.00 and ACOS norms of 25–35%. The engine underneath the growth is measurement — retail networks own the transaction, so they can tie an ad impression to an actual purchase without a single third-party cookie. Every figure below carries its source, and the full cross-channel context sits in our free Paid Media Benchmarks report.

How fast is retail media growing?

Faster than any other major line item in the media plan. eMarketer and GroupM forecasts have retail media compounding near 20% annually, which makes it the fastest-growing budget line in the industry — and the growth arrives by taking share of performance budgets rather than by expanding them. Money that used to buy open-web display and a slice of paid social is being rebooked against retailer search results and on-site placements.

Two forces drive the compounding. The first is inventory: nearly every large retailer now operates its own network, so grocery chains, big-box stores, marketplaces and delivery platforms all sell placement against their shelf traffic, with Amazon anchoring the category. The second is measurement quality, covered below — when the ad and the checkout live in the same system, the attribution argument that haunts every other channel simply gets shorter.

For operators, the planning consequence is competitive rather than optional: if your category's demand shops on retailer sites, your competitors are already bidding on it there. Where retail media sits relative to the other auctions — search as the intent engine, social as the reach engine — is mapped in our paid media statistics roundup, which compiles the CPC and CPM medians for every major channel in one place.

What does an Amazon click cost?

Amazon anchors the category, and its Sponsored Products CPCs cluster near $0.90–1.00 per Amazon agency trackers such as Adbadger. Set against the $4.66 cross-industry Google Search median from WordStream/LocalIQ, an Amazon click costs roughly a fifth as much — for a searcher who is often one click from checkout rather than three pages of research away.

The caveat that keeps the comparison honest: cheap clicks are only cheap relative to what they convert. A crowded Amazon category with three aggressive competitors can price its terms far above the tracker median, while a niche with weak competition sits below it. And the platforms answer different questions — Amazon captures shoppers who already chose the marketplace; Google Shopping reaches them a step earlier in the journey. The full head-to-head, including where each wins on margin math, is in our Amazon Ads vs Google Shopping comparison.

What is a good ACOS?

Published norms run 25–35%, and the spread inside that range is your margin structure talking. ACOS — advertising cost of sales — is ad spend divided by ad-attributed revenue, which makes it the inverse of ROAS: a 25% ACOS is a 4x ROAS, a 33% ACOS is roughly 3x. The number every operator should compute before touching a bid is break-even ACOS, and it equals contribution margin exactly.

A worked example with round numbers: a $40 product carrying a 35% contribution margin makes $14 per order after variable costs. Spend up to $14 in ads per attributed sale and the order still clears; spend more and each conversion loses money. That is an ACOS ceiling of 35% — so a category norm of 25–35% spans everything from comfortably profitable to dead break-even depending on which side of the margin table you sit. Our free ROAS & Break-Even Calculator runs this conversion both directions, turning your real margins into the exact ACOS or ROAS line your account has to beat.

Two refinements separate practiced Amazon operators from dashboard readers. First, split branded and non-brand ACOS: branded terms convert shoppers who searched for you by name, so their low ACOS flatters the blended number while telling you little about growth. Second, watch total sales alongside ad-attributed sales — ads that lift organic rank pay a dividend the per-campaign ACOS never shows, which is why many teams steer on the blended ratio once spend matures.

Why is closed-loop attribution the channel's edge?

Because the network sees the purchase. On the open web, an ad platform observes a click and then reconstructs what happened afterward through pixels, modeled conversions and attribution windows — machinery that privacy features keep degrading. A retail media network skips the reconstruction: the impression, the click and the transaction all happen inside one logged-in system. That closed loop is exactly the measurement quality the open web lost, and it is the reason procurement teams keep approving retail media budgets while other line items fight for renewal.

The discipline is remembering what the closed loop does and does not prove. It proves the ad touched a real purchase. It does not prove the purchase needed the ad — a sponsored placement on a branded search you would have won organically books revenue the business was collecting anyway. The same overlap problem shows up across every platform: summed platform-attributed revenue routinely exceeds real blended revenue, which is why blended guardrail metrics and incrementality tests remain the operator's job even on the cleanest network. Treat the closed loop as better evidence, and keep asking the incrementality question anyway.

There is also a portfolio effect worth naming: as measurement elsewhere gets harder, the channels that can prove outcomes attract budget beyond their raw performance. Part of retail media's 20% compounding is genuine shopper behavior; part is measurable-ness winning planning meetings. Both are real forces, and only one of them shows up in your contribution margin.

How should you budget for retail media in 2026?

Start from auction physics. Cross-industry CPCs have inflated at roughly 10% a year on the major platforms per WordStream's year-over-year studies, and retail media auctions face the same pressure as more brands arrive — a flat budget quietly buys less each year. Then remember the spread that medians hide: the gap between an average account and a top-quartile account on the same channel runs 2–4x in published agency datasets, and structure, targeting hygiene and creative decide most of it.

The practical sequence: defend branded terms first, since they are the cheapest revenue in the account and the first thing a competitor will raid. Expand into category terms only where your ACOS ceiling gives the auction room. Then pressure-test the whole allocation against your other channels — our free Media Mix Planner lets you model a retail media line alongside search and social against editable benchmarks, so the budget argument happens on paper before it happens in the account. This is also the weekly work of a performance media practice: finding each channel's saturation point and holding the portfolio at the profit-maximizing edge as auction prices move.

What should operators take from these numbers?

Three moves cover most situations. First, get the margin math down before the channel math: break-even ACOS from real contribution margin, then read the 25–35% norms as competitive context. Second, instrument the incrementality question early — branded versus non-brand splits and periodic holdout tests keep the closed loop honest. Third, rebenchmark quarterly, because a channel compounding at 20% a year reprices faster than annual planning cycles assume.

For the neighboring numbers, our marketing statistics library collects every sourced figure in this series: the SMS marketing statistics for the owned-channel counterweight, the landing page conversion statistics for what happens after the click, and the martech statistics for the stack that has to hold it all together.