Why Retail Media Networks Keep Winning Budgets (And Losing Trust)

Most brands still can't measure what matters. Here's what's actually working.

The Clean Room Disconnect

66% of organizations adopted data clean rooms, but the challenge lies in wiring them into planning, activation, and reporting rather than keeping them as isolated pilots. The technology works. The infrastructure doesn’t exist in integrated form.

Among organizations using clean rooms, the most common challenges include integrating with other practices and scaling as new data is collected. Yet only 63% of CMOs say clean rooms are among their top two investments, meaning 37% ignore what might be the only path to proper retail media measurement.

Budgetary buy-in is key, with adoption often lagging in marketing functions under tight budgets or anticipating constraints. Without immediate pressure, CFO focus tends toward campaign spending rather than long-term collaboration technologies.

The Retail Media Reality

Nearly two-thirds of marketers increased retail media investments in 2024 despite lackluster performance and measurement concerns. They’re throwing money at a channel they’re not entirely convinced works.

Amazon Ads is estimated larger than all other U.S. retail media networks combined, hitting $47 billion in 2023 while Walmart captured $3.4 billion. Amazon and Walmart combined will capture more than 84% of all retail media ad spending in 2025, with the share allocated to all other networks increasing less than 1 percentage point between 2019 and 2024.

Walk into a media planning meeting and ask: “What’s our retail media strategy?” You’ll get a spreadsheet showing budget allocation across twelve networks. Ask: “What are we learning from each network that we can’t get from Amazon?” You’ll get silence.

Media buyers increasingly compare Amazon’s DSP to platforms like Trade Desk or DV360, while Amazon Marketing Cloud serves as “the glue” connecting top and bottom funnel media. Amazon built actual infrastructure. Everyone else built ad units.

The Measurement Problem

Every retail media network operates as a walled garden, measuring ROI in its own silo. Spend $100K on Amazon and $100K on Target. You’ll get two separate ROAS figures—one for Amazon sales, one for Target sales. But did you sell $200K of incremental product, or did the same customer see ads on both platforms?

Attribution becomes fuzzy across retailers, with platforms often using last-click or last-touch within their ecosystem while cross-platform measurement remains challenging. A customer sees a sponsored product on Walmart.com, searches on Amazon a week later, and buys there. Walmart’s platform won’t count it. Amazon will. Neither tells the full story.

Just around one-quarter of organizations say they are proficient in measuring incrementality for retail media. Three years into the boom, and most marketers still can’t confidently answer whether their retail media spending actually drives new sales or just captures purchases that would have happened anyway.

What Actually Works

Kellanova unified postcode-level Circana data with 20 million addressable records from Experian to understand household composition and interests, identifying two distinct Special K consumer segments. The tailored approach drove a 9% sales lift with price-conscious consumers and 36% with loyalists. That’s real performance, measured properly.

But here’s the catch: this took months of work with specialized partners. The infrastructure exists in isolation. It’s not wired into planning, activation, and reporting workflows.

WPP recently acquired data clean room startup InfoSum, underscoring how central these tools have become. Agencies see the future. But most brands aren’t modeling for this reality.

The Technology Gap

Snowflake was named a Leader in the 2025 IDC MarketScape for Data Clean Rooms, with the study noting their acquisition of Samooha enhanced capabilities and simplified setup for less technical users. LiveRamp was recognized as a Leader with extensive partner network, interoperable architecture, and native activation capabilities.

Organizations want to collaborate across multiple clean rooms and vendors without sacrificing control or compliance, with emerging IAB Data Clean Room Standards aiming to make this future a reality. Real-time use cases are gaining traction, requiring dynamic campaign measurement across platforms.

CDPs focus on making a brand’s first-party data accessible but lack tools for secure collaboration with external data owners. Most brands buy audience segments from data providers, layer on some first-party information, and hope algorithms figure out the rest.

The Integration Challenge

For comprehensive customer strategy, you must integrate data from multiple sources like retail POS, OTT viewership, and social media use. Data clean rooms allow cross-channel analysis to identify shifts in consumer behavior, respond with targeted marketing, and measure customer reaction.

Theoretically. In practice, most marketers juggle dozens of platforms with incompatible data structures, limited APIs, and measurement frameworks that don’t communicate.

Kellanova’s first clean room project took months, while similar initiatives now take three to four weeks. Progress, but still far from the always-on, real-time decisioning marketing leaders imagine.

For insights to drive outcomes, Kellanova relies on agency partners who know how to unlock data and apply it to creative and media decisions. The creative team needed to change call-to-action messaging. The media team needed to shift from TikTok to Pinterest. Good data doesn’t automatically translate to good marketing.

Privacy Under Pressure

An estimated 23-56% of ad spend is currently wasted, and that’s before third-party cookies are completely deprecated. Acquiring user-level data has become challenging due to privacy regulations and changes implemented by platforms like Apple.

New regulations like EU Digital Markets Act and Data Governance Act push for greater transparency, fairness, and control over data sharing. GDPR was Europe’s first comprehensive privacy law in 2016, marking the start of a global shift toward consumer data privacy.

The trend is clear: more regulation, less access, higher technical requirements for compliant data usage. Early adopters of data clean rooms are more likely to have better data models trained on more use cases to deliver AI-powered insights, positioning marketing investments to drive value when attribution models like third-party cookies are deprecated.

Waiting for perfect clarity before investing means being two years behind competitors who started testing now.

The Practical Path

Healthcare targeting understands decision windows include filling prescriptions, switching treatments, preparing to talk with physicians—moments when people are exploring and forming questions. Campaigns connected to these signals reach consumers when influence is still possible, before decisions are locked.

The majority of recommendations tend to be driven by a small subset of professionals, so the right foundation can eliminate wasted spend by accurately identifying decision leaders. That’s genuinely valuable. But only if you can act on it faster than your competitors.

Data clean rooms improve marketing effectiveness by providing comprehensive campaign data crucial to maximizing ROAS, enabling attribution models that combine transaction data with ad performance metrics.

Kellanova is applying the clean room strategy to other brands and different markets, but doesn’t produce one-size-fits-all models since regions have varying data availability and privacy regulations. They intentionally choose more challenging markets first, ensuring if it works there, it’ll scale to other key markets.

Data clean rooms can help reduce costs and boost sales while protecting customer data privacy and maintaining compliance as part of a larger data collaboration strategy.

The winners won’t be companies with the most data. They’ll be companies that use available data most effectively to understand decision processes rather than just targeting individuals. Precision targeting was always the wrong phrase. Precision understanding is what matters.