When retail media beats paid social
Retail media and paid social answer different questions. A sponsored product ad answers a shopper who is already searching a retailer for the category. A social ad has to create the want before it can sell. Retail media tends to win for replenishment and comparison-shopped categories that are already listed, in stock and winning the buy box. Social tends to win for new, visual or considered products that shoppers do not yet search for by name. Compare the two on contribution after the retailer’s fees, not on reported return on ad spend.
Executive findings
- 1
Walmart Sponsored Search for Marketplace sellers charges per click with no setup or monthly fee, and offers Sponsored Products, Sponsored Brands and Sponsored Videos.
Documented: official source. Scope: Walmart Marketplace advertising page as published, as of . [1]
- 2
Walmart’s own guide states that advertised items must be in stock and win the buy box to serve Sponsored Products ads. The guide dates from 2020, so current eligibility should be checked in the advertiser help centre.
Documented: official source. Scope: Walmart Connect beginner guide (2020), as of . [2]
- 3
Amazon Sponsored Products are bought per click and need an eligible listing, which ties the channel to catalogue readiness.
Documented: official source. Scope: Amazon Ads Sponsored Products page, as of . [3]
- 4
In our illustrative model, a replenishment product earns 7.50 per order through retail media against 4.60 through paid social to the own site, until social ad cost per order falls below 6.10.
Illustrative model, not a benchmark. Scope: FOCUSADS worked model with labelled assumptions, not a benchmark, as of
- 5
The IAB/MRC retail media guidelines and IAB incrementality guidance set the vocabulary for asking networks how sales are attributed and whether holdouts are offered.
Secondary source. Scope: US-led industry guidelines used internationally, as of . [4]
Method and limitations
- Question
- For which product categories should a seller move budget from paid social to retail media, and how should the switch be tested?
- Written for
- Ecommerce and brand managers selling through both their own site and large retailers or marketplaces.
- Decision at stake
- How to split next quarter’s budget between retailer ads and paid social for each product line.
- Sources
- Advertiser pages and guides from Walmart Connect and Amazon Ads, the IAB/MRC retail media guidelines, IAB incrementality guidance, and Meta’s attribution settings documentation.
- Period
- Documentation as published on 17 October 2026.
- Markets
- United States examples; the method applies to any retailer that sells ads and reports sales.
- Definitions
- Search-led category: shoppers type the category or brand into the retailer’s search box before buying (batteries, filters, pet food, replacement parts).
- Discovery-led category: shoppers rarely search for the product until they see it (new brands, design-led goods, gifts).
- Contribution per order: price minus cost of goods, retailer fees, fulfilment and ad cost per order.
- Limitations
- We publish no cost-per-click or conversion-rate benchmarks; both vary by category and season and must come from your accounts.
- Retailers and Meta measure sales with different windows and methods, so their reported returns cannot be compared directly.
- Walmart’s eligibility rules are cited from a 2020 guide; confirm current rules before planning.
Where each channel finds the buyer
FOCUSADS view
A retailer’s search box collects intent that already exists. If shoppers search “HEPA filter 16x25” on a retailer, the sponsored slot next to that query is close to the buy button. Paid social works earlier: it shows the product to people who were not looking, then relies on them to come back.
That gives a simple test. Look up how often the category is searched on the retailer (most networks show query volume or impression share in their console). If the category is searched often and your product is listed, retail media should get the first test. If nobody searches for it yet, social has to build the demand first.
Worked model: contribution per order on each route
Illustrative model, not a benchmark
A replenishment product sells for 30 on both a retailer and the brand’s own site. The comparison uses contribution, not reported return.
| Line | Retail media | Paid social to own site | How it is calculated |
|---|---|---|---|
| Price | 30.00 | 30.00 | Same shelf price |
| Cost of goods | 9.00 | 9.00 | Assumption |
| Retailer referral fee or payment fee | 4.50 | 0.90 | Assumption: 15% referral against 3% payment processing |
| Fulfilment and shipping | 5.00 | 6.50 | Assumption |
| Ad cost per order | 4.00 | 9.00 | Assumption: CPC ÷ conversion rate on each route |
| Contribution per order | 7.50 | 4.60 | Price minus the four cost lines |
Assumptions
- Every figure is a placeholder; use your fee schedule, fulfilment costs and account data.
- The model ignores repeat purchases. A customer won on your own site can be emailed; a retailer customer usually cannot.
- If social ad cost per order falls below 6.10, the own-site route earns more per order in this model.
When social still wins
FOCUSADS view
Social should keep budget when the product needs to be shown before it is understood, when the margin on your own site is much higher than through the retailer, or when the repeat-purchase value of an owned customer is large. Our article on marketplace and owned-site economics works through that trade-off in detail.
What we would test, in order
| Step | Measured by | Counts as success when |
|---|---|---|
| Classify each product line as search-led or discovery-led using retailer query data and your own site search. | Retailer query impressions for the category; share of own-site sales from branded search. | Each line has a written classification before budget moves. |
| For one search-led line, move a fixed share of social budget to sponsored products for six weeks. | Total contribution across both routes, not each channel’s reported return. | Combined contribution rises against the previous six weeks, adjusted for season. |
| Ask the network for a holdout or matched-market read on the advertised products. | Incremental retailer sales, net of retailer margin. | Incremental contribution covers media cost. |
Category check: retail media or social first
The more rows that point one way, the stronger the case for testing that channel first.
| Signal | Points to retail media | Points to paid social |
|---|---|---|
| Shoppers search the category on the retailer | Often | Rarely |
| Product listed, in stock, eligible for ads | Yes | Not listed |
| Margin through the retailer after fees | Close to own site | Much lower than own site |
| Value of owning the customer relationship | Low (one-off purchase) | High (subscription, repeat) |
| Creative needed to explain the product | Little | A lot |
How this connects to our method: Start from the economics
References
- Walmart Sponsored Search for Marketplace sellers, Walmart Marketplace, accessed
- A Beginner's Guide to Walmart Sponsored Products (published 2020), Walmart Connect, accessed
- Sponsored Products, Amazon Ads, accessed
- IAB/MRC Retail Media Measurement Guidelines, IAB and Media Rating Council, accessed
- Guidelines for Incremental Measurement in Commerce Media, IAB, accessed
- About attribution models and attribution settings, Meta Business Help Center, accessed