When retail media gives the retailer more margin than selling the actual product, paid placement stops being optional for either side of the table. For the retailer it's a P&L lever. For the brand it's a recurring charge dressed up as a campaign. The CFO's question hasn't changed: "is this investment making us win in the market?" — and only two lenses, used together, can answer it.

The retailer's P&L flipped, and nobody is letting go

The number to read isn't retail media spend. It's the fraction of the retailer's operating profit that already comes from ad sales, not from selling food or products.

Scott Sanders pulls the 2025 figures: Amazon's advertising business generated USD 68.6 billion in 2025 — more than the total revenue of Publix or Ahold Delhaize USA. Walmart closed near USD 6.4 billion, growing 46%. Target Roundel approached one billion. But the number that changes the conversation is the mix: advertising is now 21% of Walmart's operating profit, ~18% of Target's, and 86% of Amazon's.

Traditional retail margin: single-digit, pressed by costs, tariffs and logistics. Retail media margin: very high, scales with volume, near-zero marginal cost. When the highest-margin segment of the retailer is the ad inventory it sells to its own suppliers, that retailer has already made the decision for you: digital space is going to be monetized more, not less. ET Edge calls this "rent" — rent, infrastructure, structural cost, and the analogy is fair because it carries steady cost while enabling scale.

Bain projected — and reality confirmed — that most retailer profit growth between 2024 and 2030 will come from "beyond traditional trade," with retail media the cleanest lever to expand that share. For the retailer, that's survival. For the brand, it's an additional annual invoice.

The brand no longer decides if it participates — only how

The financial signal is brutal: 36% of US retail media spend by brands comes out of trade marketing budgets, and another 26% comes out of shopper marketing. It isn't incremental money. It's dollars that already sat on the retailer's commercial table, now redirected into the retailer's ad server. "Paid retail media" jumped from experimental line item to 39% of CPG ad budgets in 2025, up from under 25% three years ago. 70% of CPG brands plan to increase that share further in the next 12 months (Skai 2026).

If your digital visibility depends increasingly on sponsored slots — and retailers have a structural incentive for it to depend more — then participating isn't a campaign decision, it's a condition of presence on the digital shelf. It's the 2026 version of the 90s slotting fee: if you want eye-level real estate, you pay. What changed is that the charge is now visible, measurable, and re-auctioned week to week.

For the brand, the operational consequence is one of calendar, not campaign: retail media enters the annual joint business plan with the retailer, not the quarterly media plan. And because it enters as a commercial cost, the conversation is no longer "how much do we spend?" but "is this level of spend leaving a healthier customer than before?"

The CFO's question didn't change: is this investment making us win in the market?

The CFO doesn't ask about ROAS. The CFO asks two different questions — and confusing them is exactly why half of brands say they can't prove the ROI of their retail media (Skai 2026).

Question 1: would these sales have happened anyway? This is incrementality. It's a counterfactual question and only a model answers it: ideally a geo holdout, in its absence a clean observational multivariate regression with the critical confounders accounted for. The branded-paid-search cannibalization documented by Blake, Nosko & Tadelis (2015) at eBay remains the canonical case of spend whose incrementality is essentially zero — and no ROAS will surface it. Only 21% of brands rate themselves "good" at measuring incrementality; true iROAS varies between 253% and 1,609% on the same reported spend, which gives a sense of the error size if you stop at the retailer's dashboard.

Question 2: are we gaining position in the market? This is share / competitive position. Not the same question. Your retail media can be perfectly incremental — those sales would not have happened without the spend — and you can still be losing ground because competitors are growing faster, attacking keywords where you used to be organic-strong, or buying placements on top of a higher PDP floor than yours. Incrementality answers "is the spend worth its cost?". Position answers "are we better off on the digital shelf than we were last quarter?".

Two different lenses. Confusing them — using ROAS for one, or organic position for the other — is what produces the Q4 deck where the retail media team says "we're up 30%" and the CFO says "and our share dropped two points."

The honest stack: two layers, two tools

If paid placement is now a structural cost, the discipline around how you evaluate it shifts from "best practice" to governance of recurring spend. What the brand needs on the table is a two-layer answer:

Layer 1 — Incrementality. This is where the attribution model lives. It can be an in-house MMM, geo holdouts coordinated with the retailer, or observational regression against the six variables that move rank (the honest measurement guide and the 4 confounders only scraping detects cover the methodology in detail). The vendor doesn't have to be ePerfectStore — it can be your analytics team, your MMM consultancy, or the retailer itself if you trust their closed-loop. The output is one number: an incrementality coefficient per channel, with confidence interval and documented bias, aligned with IAB/MRC guidelines.

Layer 2 — Competitive position on the digital shelf. This is where the weighted organic position thermometer lives: your weekly rank by keyword × retailer, your paid SOV vs. the competitor's, the availability timeline by SKU, and the continuous content score. It's the layer where ePerfectStore.com is the natural tool in LATAM because no retailer hands it to you and no agency measures it at the granularity needed for Mercado Libre, Éxito, Jumbo, Olímpica or Falabella simultaneously. It's the reading that answers "is this spend building presence, or only holding it?".

Without Layer 1, you don't know if the spend is worth it. Without Layer 2, you don't know if you're winning. Having only one of the two is the recipe for spending more every year on a structural cost without being able to tell the CFO with a straight face that you're better off.

A practical note: the two layers don't merge into one dashboard. Incrementality is a statistical coefficient; competitive position is a per-keyword time series. They are different data products. The discipline is reading them together, not fusing them. In the quarterly finance review the sentence is: "spend is 87% incremental (Layer 1) and our weighted organic position improved from 4.2 to 3.6 across the top 30 keywords (Layer 2). Conclusion: we're paying a cost that pays back, and we're gaining shelf." Or its opposite, if the opposite is true.

What changes at the table when you accept it's a structural cost

Three operational consequences:

  1. Retail media budget leaves tactical media planning and enters the annual joint business plan with the retailer. If it's 21% of the retailer's operating profit, that conversation is already happening at a different level. The brand still treating it as quarterly media spend will be captured in the auction without strategic counterparty.
  2. The two questions get reported separately and to different roles. Incrementality goes up to CMO/CFO. Competitive position goes up to the commercial lead and the retailer's KAM. Collapsing them into one number (typically ROAS) is what produces contradictory decisions across quarters.
  3. Investment consolidation becomes inevitable. Bain reported that 50% of US advertisers plan to cut the number of networks they work with to six or fewer. In LATAM the ceiling will be lower — 3 or 4 — because the universe of large retailers with their own network is countable on one hand. The brand that walks into that conversation without the two layers loses negotiation power.

Close

Retail media stopped being a marketing decision. It's a recurring cost line on the commercial side, defended by the math of the retailer's P&L and the structural position of the digital shopper. Fighting it is not the play. Reading it is.

The brand that survives this cycle is the one that shows up to every conversation with two answers, not one ROAS: "this spend is X% incremental, with this bias, per this model" (Layer 1) and "our weighted shelf position moved Y points across these categories, against Z competitor moves" (Layer 2). Any combination of those two answers — even a bad one — is actionable. ROAS alone is not.

ePerfectStore.com lives in Layer 2: continuous organic-position thermometer, competitive paid SOV, SKU-level availability, and the content score the retailer doesn't hand you. Your attribution model can live elsewhere. Living separately is fine. What can't happen is that they live missing.

Sources

Is your retail media spend growing every year but you still can't tell the CFO if you're better off? ePerfectStore.com is Layer 2 of the honest stack: continuous organic-position thermometer, competitor SOV, availability, and content score.

Back to Blog