The “safest” spend in 2025 is quietly taxing your brand. If your top-performing channel is retail media or OTA ads, chances are you’re buying back customers you would have converted anyway. Retail media incrementality is the share of sales caused by retail media ads that would not have happened otherwise; it isolates true lift by excluding cannibalized branded traffic, repeat buyers, and baseline demand. It is distinct from closed-loop attribution, which merely ties conversions to ad exposures inside a walled garden without proving those sales were additive.
The Association of National Advertisers’ 2024 Retail Media Measurement Guidelines called out inconsistent definitions, opaque reporting, and the risk of conflating attributed sales with incremental impact—especially where closed-loop platforms can’t account for cannibalization or branded traffic (ANA, 2024). Meanwhile, Insider Intelligence projects continued double‑digit growth of retail media ad spend into 2025 as third‑party cookies erode and marketers chase “provable” outcomes inside walled gardens (Insider Intelligence/eMarketer, 2025). That demand is rational—but the measurement is not.
RETAIL MEDIA INCREMENTALITY, NOT ROAS
A high platform-reported ROAS is not evidence of growth; it’s evidence of targeting intent. In fashion, lifestyle, and hospitality, marketplaces and OTAs rank and advertise on your brand terms, creating gorgeous ROAS dashboards while quietly eroding margin and masking brand weakness. The ANA’s 2024 guidelines explicitly warn that closed‑loop attribution can double-count or misattribute sales when platforms fail to control for cannibalization and brand-seeking traffic. This is the difference between credit and causality.
There’s precedent: eBay’s landmark randomized experiments found that paid search on brand keywords delivered little to no incremental sales when organic results were strong, effectively paying for clicks that would have occurred anyway (Blake, Nosko, and Tadelis, 2015). Today’s retail media behaves similarly: brand-term capture and heavy retargeting look efficient but often harvest demand you already own. When ROAS is highest where intent is highest, you’re measuring intent—not impact.
HOW TO MEASURE TRUE LIFT IN 2025
If it doesn’t move baselines, it isn’t growth. Start with geo-experiments: hold out matched markets from retail media or brand-term capture for 4–8 weeks, then compare sales deltas vs. controls to isolate incrementality. Complement this with media mix modeling (MMM) to quantify cross-channel effects and calibrate with retailer sales logs. The ANA urges standard definitions and lift-based testing over pure attribution, and these methods deliver exactly that (ANA, 2024).
Set buying guardrails before you test. Cap spend on brand-term capture (especially your own name plus “coupon” or “sale”), throttle retargeting frequency, add negative keywords for navigational queries, and separate prospecting budgets from retention. When you migrate brand-term dollars into true prospecting and see flat or rising revenue in holdouts, you’ve found incremental headroom. Growth is the part that survives a holdout.
STOP PAYING THE TAX: BUILD OWNED DEMAND
The antidote to retail media rent is owned demand creation. Reweight toward brand equity, distinctive creative, and direct relationships that marketplaces can’t tax. On Shopify Plus, that means building first‑party data through on‑site value exchanges, then compounding it with Klaviyo flows and predictive segmentation to lift repeat purchase without paying a walled garden for your own customers. You’re shifting from capture to creation.
Follow the money signals. Amazon reported advertising revenue growing more than 20% year over year in 2024 (Amazon filings), mirroring Insider Intelligence’s double‑digit growth outlook for retail media into 2025. That spend will keep flowing until more brands run lift tests, cut non‑incremental capture, and reinvest into brand and lifecycle. The most defensible ROAS is the one no retailer can charge you for.
The fastest way to stop the leak: measure lift, cap capture, and invest in brand so you never have to buy back demand you already own.
The short answer: retail media isn’t your safest 2025 bet unless it proves incrementality in controlled tests. A resilient measurement strategy in 2025 is to quantify true lift with geo‑experiments and MMM, cap spend on brand‑term capture, and reweight budgets toward brand equity and owned demand systems on platforms like Shopify and Klaviyo—work EDEUS Studio builds end‑to‑end.
