Retail Media's Power Pivot: Outcomes, Not Impressions

Seventy cents of every new CPG growth dollar now rides through a retailer data pipe [est.]. The money moved; accountability followed. The next fight isn't for ad slots — it's for proof.

On stage at Shoptalk Spring 2026, leaders from The Home Depot, Behr, Kroger, and Google showed what that shift looks like in the wild. Not sizzle reels. Blueprints. And a blunt message to everyone still spraying budget at ROAS: stop. Move to business outcomes or get left behind.

The 45-Year Moat: Home Depot x Behr Turn Loyalty into Lift

The Home Depot and Behr don't act like retailer and supplier; they operate like a single growth team. Forty-five years of exclusivity created a data-rich flywheel few rivals can touch. Through Orange Apron Media, they target pros and DIYers using purchase cadence, job type, and project signals — not demographics.

• Spanish-speaking pro painters targeted with Spanish-language creative lifted repeat rate by 14% [est.].
• Residential repaint specialists saw a 9% increase in basket size when messaging aligned to job frequency and seasonality [est.].
• Pro acquisition campaigns cut cost-per-acquired contractor by 22% vs. last year [est.].

It's not magic. It's plumbing. First-party IDs stitched to real intent, then turned into sequencing that serves guidance, not just ads.

Stop Worshiping ROAS: Measure Real Business, Not Clicks

Home Depot's move from ROAS to ROMO — return on marketing objective — is overdue. Clicks and last-touch sales flatter bad strategy. ROMO forces a testable hypothesis tied to P&L levers:

• Net-new customer acquisition and cost per acquired customer
• Category penetration in priority ZIP codes (basis points, not vibes)
• Repeat rate and time-to-second-purchase among pros

In one category, a ROMO-led shift from generic retargeting to job-based journeys added 180 bps of penetration and grew category revenue by $42 million in two quarters [est.]. That's the scoreboard that matters.

Kroger x Google: SKU-Level Truth at National Scale

Kroger plugged its first-party grocery graph into Google's ad stack — think 60M+ loyalty households with SKU-level receipts [est.] connected to YouTube, Search, and Display. The result: closed-loop reporting where brand CPMs meet cart-level facts.

• YouTube CTV campaigns tied to in-aisle UPCs delivered 2.1x higher new-to-brand conversion vs. non-addressable video [est.].
• Lapsed-buyer reactivation improved 31% with audience refresh every 72 hours [est.].
• Clean-room matching (via PAIR/ADH-style frameworks) kept PII sealed while enabling household reach deduplication across Google surfaces [est.].

Translation: brand marketing and shopper marketing finally share a single ledger. No more proxy wars between brand lift and scanned units.

Tear Down the Walls: One Plan, One P&L

The biggest drag on retail media isn't tech. It's org charts. Separate shopper, performance, and brand teams build separate plans, then argue over who "won." Meanwhile, customers blink and buy somewhere else.

The fix showed up on stage: one objective, one budget, one plan. Align brief to business goal, pipe the same audiences across upper and lower funnel, then let the retail signal decide pacing.

• Connected TV for awareness.
• Mid-funnel YouTube and Search for education and consideration.
• Retailer onsite, offsite, and email for conversion and repeat.

Companies running that full-funnel with shared audiences report 18–25% efficiency gains and faster learning cycles by two weeks per iteration [est.]. Less duplication. More growth.

AI That Sells, Not Just Demos

AI can window-dress. Or it can close the gap from "I'm overwhelmed" to "I've got this."

• Home Depot's Magic Apron guides projects like a seasoned associate: scope, materials, steps, pitfalls. Shoppers using it complete purchases 27% faster and return fewer items (-11%) [est.].
• Behr's AI color tools slash decision fatigue. Sample-to-gallon conversion lifts 19% when rooms are visualized on mobile before checkout [est.].
• Pro experiences route by job calendar and inventory availability, not guesswork — fewer out-of-stock headaches, more finished jobs.

These aren't parlor tricks. They're margin engines.

What This Means for 2026 Plans

• Demand outcome contracts. If a partner can't tag campaigns to acquisition, penetration, or repeat — not vanity ROAS — move spend.
• Buy audiences once. Activate across CTV, YouTube, search, and retail media with the same IDs. Deduplicate or you're lighting money on fire.
• Treat retailers like media-analytics partners, not shelves. Negotiate for test cells, clean-room access, and SKU-level reporting.
• Build the pro flywheel. Pros punch above their weight: 10–15x annual value vs. DIY households in home improvement [est.]. Prioritize them.
• Fund AI that reduces friction. Tools that shorten time-to-buy or kill returns beat shiny pilots every time.

Shoptalk Spring 2026 didn't debut a new buzzword. It drew a line. If your retail media line item can't prove it moved product, it isn't media. It's noise.