Activating retail media and shopper marketing for measurable growth

# Activating retail media and shopper marketing for measurable growth

A brand manager opens two dashboards on Monday morning. On the left, Amazon Ads shows Sponsored Products spend, click share, and an ACOS (Advertising Cost of Sales: ad spend divided by the sales those ads drove) drifting past target. On the right, Walmart Connect reports sales lift among households that saw a digital offer, matched to what those households actually scanned at checkout.

Same brand, same week, two incompatible stories about what drove growth.

This lesson is about reading those dashboards, building campaigns that close the loop from exposure to basket, and proving which dollars were incremental: sales that would not have happened without the campaign.

Why retail media changed FMCG marketing

A retail media network (RMN) is a retailer selling advertising on its own properties (site, app, in-store screens and audio), using its first-party purchase data both to target the ad and to report the outcome.

The difference from broadcast is the receipt. A TV plan models exposure. Amazon or Walmart can tie an impressionimpressionThe total number of times an ad or piece of content is displayed, regardless of clicks. Each display counts as one impression, even to the same person.View full definition → to a specific basket, by customer ID, within days.

Two structural facts make this bite harder in FMCG than in most sectors. Baskets are small and repeated, so purchase data is dense and refreshes weekly rather than annually. And a growing share of grocery discovery now starts in a retailer's search bar, even on trips that finish in a store.

US retail media spend is widely estimated above $50 billion a year, and Amazon alone reported advertising services revenue north of $45 billion in 2023. Treat market totals as estimates; the direction of travel is not in doubt.

The two dashboards, decoded

Amazon: the digital shelf

On Amazon your product competes inside a search results page. The levers:

• Sponsored Products and Sponsored Brands: paid placements bought against keywords like "electrolyte drink."
• Share of shelf: the percentage of relevant results your items occupy, organic plus paid.
• Retail readiness: in stock, strong title, images, ratings holding above roughly 4 stars. Traffic sent to a weak page is money burned.

Watch ACOS alongside TACOS (total ad spend against total sales, organic included). A flattering ACOS on one campaign can sit inside bloated total spend. For anything beyond the packaged report, Amazon Marketing Cloud lets you query exposure and conversion signals yourself, which is the only way to see overlap between campaigns.

Walmart: the loyalty-data shelf

Walmart Connect sells on-site search and display plus in-store inventory: self-checkout screens, the TV wall, in-store radio across thousands of US stores. Walmart's 2024 acquisition of Vizio, at around $2.3 billion, was openly about adding screens and viewing data to that business. You can target lapsed buyers of your brand, or households currently buying a competitor, and receive lift among an exposed group against a matched control.

The in-store side carries an edge case worth planning for. A self-checkout screen produces no click and no cookie, so measurement falls back to matched store panels: exposed stores against control stores, read over weeks, with store-level noise (weather, a rival's display, a stockout) large enough to swamp a 2% lift.

Alibaba: buying the funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition → by stage

Alibaba's Brand Databank counts consumers by AIPL stage (awareness, interest, purchase, loyalty) and shows how many crossed between stages in a period. That reframes the media buy. Instead of cost per clickcost per clickCost Per Click (CPC) is the average amount you pay each time someone clicks your ad. It is a core pricing metric for paid search and social advertising.View full definition →, the question becomes what it cost to move 100,000 consumers from interest to first purchase, and what share of them bought again within 90 days. Few Western networks report this cleanly yet, and it is the sharpest available definition of what shopper media is supposed to do.

Retailers publish their own specifications; the IAB's retail media resources collect them, with the caveat that IAB members include the networks selling the inventory.

Designing a closed-loop shopper campaign

Shopper marketing means activity aimed at converting a shopper close to the moment of buying. Retail media is now its main engine.

1. Define the audience by behavior, not demographics.

Not "women 25 to 44" but "households that bought the category and not our brand in the last 90 days."

2. Match the message to intent.

• High intent (searching "pasta sauce"): Sponsored Product plus an offer.
• Low intent (browsing recipes): display, working on awareness.

3. Bridge digital and physical.

Offer clipped to the loyalty account, redeemed at the till. Online exposure, offline purchase, reconciled by ID.

4. Fix the measurement plan before launch.

Choose the control group and the success metric now. Retrofitting a control group after the fact is not possible.

🎬 [VIDEO: "How Retail Media Networks Work" - youtube.com - a clear explainer on how retailers monetize first-party shopper data]

Negotiating the joint business plan

A Joint Business Plan (JBP) is the annual agreement between a brand and a retailer setting shared growth targets, promotional calendar and, now, media commitments.

Historically the JBP arbitrated the trade investment the P&L lesson deals with. Media has since moved into the same meeting, which changes who is in the room: the retailer's media sales team has its own revenue target, separate from the buyer's margin target, and the two will not always ask you for the same thing.

Carrefour Links, built with Publicis in 2021, sells targeting and measurement off Carrefour's loyalty data across its European and Latin American markets. It also shows the practical limit: consent regimes, loyalty penetration and reporting formats differ country by country, so a "pan-European" retail media plan is usually a stack of incomparable national reports rather than one number.

Principles for a stronger JBP:

• Tie media to mutual growth. Anchor commitments to category growth; that is what the buyer is measured on.
• Agree the measurement standard upfront. Incrementality methodology, control group definition, and whether you get exposure-level data or a summary PDF.
• Separate media from listing. Once ad spend becomes an unwritten condition of shelf space, you have lost the ability to switch it off when iROAS falls.
• Resist the always-on offer. Constant depth trains shoppers to wait, and the damage outlives the campaign.

A useful question: "If I move budget here, what specific in-store or digital-shelf outcome do you commit to?"

Measuring incrementality (the part that earns the budget)

Most retail media reporting shows attributed sales: any purchase by someone who saw or clicked. That overstates impact, because some of those shoppers were buying anyway.

The test-and-control method

1. Randomly split the target audience: one group eligible for the ad, one held out.

2. Run the campaign.

3. Compare purchase rates.

Incremental sales = test group purchases minus control group purchases.

iROAS = incremental sales revenue / ad spend

Example (illustrative numbers):
  Test group sales:      $500,000
  Control group sales:   $420,000
  Incremental sales:     $ 80,000
  Ad spend:              $ 40,000
  iROAS = 80,000 / 40,000 = 2.0

An iROAS of 2.0 means every dollar drove two dollars of additional sales. Compare that to gross margingross marginGross margin is the share of revenue left after subtracting the direct cost of producing goods or services, expressed as a percentage of revenue.View full definition → before calling it profitable. High attributed ROASROASReturn on Ad Spend (ROAS) measures the revenue generated for every unit of currency spent on advertising, calculated as revenue divided by ad cost.View full definition → with low iROAS is the classic tell: you paid to reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → buyers you already had.

Two cautions the reports rarely print. First, statistical power. If your brand is bought by 3% of households in a given month, detecting a 5% lift needs exposed and control cells in the hundreds of thousands; most two-week campaigns are simply too small to read, and the "lift" reported is noise dressed as a result. Second, the scope of the word incremental. A Walmart test measures incremental sales at Walmart. If the shopper would have bought your brand at another banner that week, the retailer's number is real and your brand-level number is smaller. Cross-retailer panel data is the only check.

Reading the dashboard like an operator

Separate defend from grow. Branded keyword campaigns show excellent ACOS and poor incrementality: you are paying to appear in front of people typing your name. Some defense is rational when competitors bid on your brand; funding it as growth is not.

Watch the halo, but do not assume it. Digital-shelf spend can lift in-store sales. Believe it when a control group says so.

Reconcile the two dashboards. Amazon and Walmart can move in opposite directions in the same week because of a promotion, a stockout or a seasonal skew. Investigate divergence instead of averaging it away.

A simple weekly operating rhythm

• Monday: iROAS and TACOS by retailer. Flag anything below the margin threshold.
• Midweek: retail readiness (stock, ratings, content) before releasing more spend.
• Monthly: results into the JBP conversation with each retailer.

Common failure modes

• Spending behind out-of-stock lines. The fastest way to fund a competitor's conversion.
• Confusing correlation with cause. Sales rose, and so did the season, the display and a rival's price.
• Letting the seller grade its own homework. Every network reports on the media it sold you.
• Burning a narrow audience. Retarget a 200,000-household segment hard enough and frequency climbs while reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → stalls; costs rise, lift does not.
• Booking media as the price of shelf. It becomes untouchable, then permanent.

Key takeaways

• Retail media works because the retailer owns the receipt, closing the loop from exposure to basket in a way broadcast cannot.
• Target by behavior, lapsed and competitor buyers, and bridge exposure to the till with loyalty-linked offers.
• Demand incrementality, not attribution, and check that the test had enough buyers to measure anything.
• Retailer-incremental is not brand-incremental. Sales moved from one banner to another still look like growth on one dashboard.
• Treat the JBP as a growth pact, with agreed measurement, and keep media spend separable from shelf access.