"Retail media networks shouldn't sell 'awareness' or 'impressions' or 'conversion,'" Gray wrote on the social media platform in October. "They should sell reach, context, salience and physical availability because these are the things that deliver brand growth."
The post ignited something. Dozens of responses poured in from retail media leaders, agency executives, measurement specialists, and brand marketers.
What became clear wasn't that one side was right and the other wrong. It's that retail media has reached a crossroads where theoretical ideals about how brands grow are colliding with operational realities about how marketing budgets actually get allocated.
The 'How Brands Grow' gospel meets retail media
Gray's argument is grounded in the Ehrenberg-Bass Institute's research, popularized in the 2010 book 'How Brands Grow' by Byron Sharp. The core principles: brands grow through penetration (reaching more buyers), not by increasing loyalty among existing customers. Mental availability (being easy to think of) and physical availability (being easy to buy) matter more than persuasion or conversion optimization.
"When shoppers need or want something within a particular category, they buy something that does the job, is easy to think of and easy to buy," Gray wrote in his post. "Brands compete on mental and physical availability, not impressions and clicks."
His provocation: retail media is uniquely positioned to deliver both mental and physical availability at the moment shoppers enter a category. It's "the most contextually powerful channel for refreshing memory structures at the moment of category entry," he argues. A shopper searching for Greek yogurt on Walmart.com is literally in the market right now. That context makes retail media worth a premium.
When I asked Gray what metrics he'd propose instead of ROAS and clicks, he pointed me toward penetration and new user growth -- measures that brands should track "not just at a market level, but within retailers." He acknowledged these measures will take time to become mainstream. Despite the many endorsements of this concept, when I pressed for examples of retailers or solutions leading this charge, I wasn't given any specifics.
That non-answer is telling. Gray's thesis is sound -- brands do grow through penetration and availability. But translating philosophy into actionable measurement frameworks that brands can actually use to allocate budgets is where things get murky.
Why retail media metrics are broken
Gray isn't alone in arguing that the current state of measurement is flawed. Jordan Witmer, managing director of retail and commerce media at agency Salt XC, laid out exactly how current systems distort retail media's value.
"Retail media gets crushed in impression-based modeling approaches," Witmer told me. Because RMN budgets are usually steady week to week, and because the channel generates fewer impressions, the better it performs, MMMs often read it as under-delivering. Models like variation. Retail media doesn't provide much. Meanwhile, channels that pump out impressions cheaply (eg social and open-web display) look great on paper even when retail media is quietly driving more real conversions.
Yuni Baker-Saito, chief executive of ad tech company Chicory, takes the critique further. "Marketing has moved too far toward data-driven decision-making," he argues. "Intuition and context have to matter again."
Baker-Saito's thesis: marketers have become so dependent on measurement that they've lost sight of context and human behavior. Most measurement systems -- MMM, closed-loop attribution, programmatic models -- have inherent distortions that over-reward impression volume and scale. "Decision-making has become driven by what is legible rather than what is meaningful," he says.
Chicory's model is built on the idea that context matters more than attribution math. Someone viewing a recipe for Greek yogurt is already in a high-intent mindset. Adding retail media data -- targeting recipe viewers who haven't bought Greek yogurt in two weeks -- creates relevance that raw metrics can't capture.
The funnel isn't dead, But control is
Anne Hallock, vice-president of sales at Mirakl Ads, offers a reframe that bridges Gray's philosophy with operational reality.
"The purchase funnel isn't dead -- awareness must precede consideration, and consideration still drives preference," Hallock told me. "What's dead is the manner in which brands used to be able to control the customer journey."
She describes what agency executive Bret Starr calls in his book a "chaos pendulum": shoppers ping-ponging between social proof, search, reviews, retailer sites, and physical shelves. A shopper might discover a product through word of mouth, encounter it via sponsored search, then convert only after comparing options on a message board. What part of that journey did the brand actually influence?
This chaos doesn't mean awareness and consideration don't exist. It means brands can't dictate the sequence or measure it cleanly. Retail media's appeal, in Hallock's view, comes from high intent: "I show you an Apple ad in the New York Times, and you may or may not register it. I show you an Apple ad when you're shopping for a laptop on Best Buy.com and you will absolutely have added Apple to your consideration set."
Hallock's framing matters bridges both sides of this debate. Yes, mental availability precedes purchase. Yes, context matters. But shoppers are chaotic, journeys are fragmented, and brands need metrics that account for that chaos.
Eleanor Hayden, founder of agency Veriti, agrees with Gray that mental and physical availability matter -- but only up to a point. In her view, reach and salience are necessary, but they're not sufficient in categories where shoppers increasingly buy based on values, provenance, or quality cues. "This idea that when shoppers need something, they buy something that does the job -- that makes sense for a commodity, but not for CPG brands building engaging, values-aligned brands," she told me. "Shoppers' expectations are higher now. You have to be more than something that 'does the job.'"
That pesky budget spoils all the fun
Michelle Dooley, principal consultant at Catalyst Media Consulting, shares a simple truth: "At the end of the day, everyone in retail is looking to sell something -- and the investments made in advertising are in service of driving more sales."
Retail media started as a performance channel because retailers had data to prove ROI. Before ROAS became standard, brands ran test-and-control groups with public service announcement ads as controls. Getting to statistical significance with channel-level budgets was hard. ROAS became the metric that was achievable, relatively easy to calculate, and within the retailer's control.
"While ROAS has its limitations, different buyers will care about different things," Dooley notes. "If the buyer is a performance buyer, we need to talk their language. If the buyer is a brand buyer, we need to speak their language -- which retail media doesn't do a great job of today."
Baker-Saito highlights another layer: which budget 'bucket' retail media spend comes from dictates what performance must look like. For years, retailers benefited from ambiguity around whether spend should come from trade marketing, shopper marketing, or national brand budgets. That ambiguity is reaching a breaking point. Retailers want national dollars because they are truly incremental 'new revenue' to them.
But brands are facing growing pressure through joint value planning sessions, spend commitments, and multi-tactic packages bundling media with traditional trade vehicles. "The source of dollars shapes expectations," Baker-Saito says. Trade dollars have always carried performance expectations tied to specific retailer outcomes. National brand budgets follow different rules -- they flow to channels that prove they drive growth, period.
Philosophy is hard to sell
The current reality is that for better or worse, retail media networks positioned themselves as performance channels. They sold ROAS. They pitched closed-loop measurement and conversion attribution -- unique attributes that only retailers can promise. So brands hired performance buyers to manage these relationships, set quarterly targets, and justify spend in budget reviews.
If those same networks are required to sell salience and mental availability instead, that's an entirely different value proposition.
Baker-Saito offers a path forward that doesn't entirely abandon performance: retail media networks should evolve toward full-funnel solutions, particularly for mid- and long-tail brands lacking access to sophisticated demand-side platforms, advanced targeting, and high-quality data. Expanding offerings -- connected TV, influencer partnerships, social integrations -- isn't about suddenly prioritizing awareness over conversions. It's about making retail media networks an accessible entry point to capabilities smaller brands couldn't otherwise afford.
Tier-one CPG companies will use retail media differently, feeding data into marketing mix models to rationalize spend across channels. But those brands have analytics infrastructure most don't.
The heart of the matter
The shoe-dragging over common retail media measurement standards shows that the industry doesn't truly believe in perfect measurement. Brands are allocating budgets right now, using the flawed metrics available. Retail media networks are building businesses right now, selling what buyers understand how to buy. And consultants are correctly pointing out that the frameworks guiding these decisions are fundamentally broken.
Gray's provocation was a hit because it forces the industry to admit what everyone already knows: ROAS optimization isn't building brands, and last-click attribution isn't telling the truth about what's working. But until someone builds measurement infrastructure that captures salience, penetration, and mental availability in ways that CFOs will approve and media buyers can action, brands will keep using the metrics we have now.