Key Takeaways
- Most CPG brands are spreading retail media budgets across too many networks and measuring none of them well.
- Last-click ROAS is the wrong primary metric for retail media networks because it rewards proximity to purchase, not actual sales influence.
- Incrementality testing is the only way to know which retail media networks are growing your brand versus cannibalizing organic sales.
- Retail media consolidation at the holding company level signals category maturity, which means CPG brands can no longer afford an ad hoc network strategy.
- The brands winning in retail media right now are running tighter network rosters with deeper investment, not wider coverage with thin budgets.
- Network-owned measurement tools have a structural conflict of interest and should never be your only source of truth.
—
Retail media networks are not all created equal, and right now most CPG brands are treating them like they are. The consolidation happening at the agency holding company level is a signal worth paying attention to: retail media has moved from experimental line item to a core media channel that demands real strategy, real measurement, and real accountability from every network in your mix.
How CPG Brands Are Currently Evaluating Retail Media Networks
The typical CPG approach to retail media looks like this: activate on the networks your retail partners push hardest, optimize toward reported ROAS, and call it a win if the numbers look good in the dashboard. The problem is that those dashboards are built by the networks themselves.
Every major retail media network, from Amazon to Walmart Connect to Kroger Precision Marketing, has a financial incentive to show you performance metrics that justify renewal. That does not make the data fraudulent. It makes it incomplete. And incomplete measurement is how brands end up spending millions on retail media that is primarily capturing purchases that would have happened anyway.
What Incrementality Actually Means in Retail Media
Incrementality is the measure of sales that would not have occurred without your advertising. It is not a ROAS number. It is not a return on ad spend calculation from a last-touch attribution window.
True incrementality testing in retail media requires holding out a portion of your audience or geography from exposure, then comparing actual purchase behavior between the exposed and unexposed groups. Very few brands do this consistently. Even fewer hold the networks accountable to it.
We have written about this problem in a broader sense before. If you want the foundational view, our piece on retail media strategy for CPG brands beyond sponsored listings breaks down why sponsored listings alone are not a growth strategy.
Evaluating Which Retail Media Networks Deserve Your Budget
Here is a direct framework CPG brands should run every quarter. It is not complicated. Most brands just never do it.
Step 1: Audit your current network roster
List every retail media network you are active on. For each one, answer three questions:
- What percentage of my category’s volume flows through this retailer?
- Can I run a third-party verified incrementality test on this network?
- Do I have access to shopper data that is actually useful for targeting, not just reporting?
If the answer to questions two or three is no, that network belongs on a probationary budget until it is yes.
Step 2: Rank networks by strategic value, not just reported performance
Reported ROAS from network dashboards is a starting point, not a conclusion. Layer on these factors:
- Retailer’s share of your category’s total sales volume
- Data transparency and access to clean, exportable audience signals
- Availability of closed-loop measurement with independent verification
- Ability to run full-funnel formats beyond sponsored product listings
Step 3: Concentrate budget on fewer networks with higher accountability
Spreading thin across eight networks to maintain retail relationships is a political strategy, not a media strategy. The brands we see outperforming in retail media are typically deep in two or three networks with structured measurement programs, not present on every network with minimal commitment.
The Conflict of Interest Every CPG Brand Needs to Acknowledge
Retail media networks sell the ads. They also measure the ads. This is not a conspiracy. It is just a structure that requires you to bring independent measurement to the table. Third-party solutions and clean room environments exist precisely for this reason.
This is also why agency consolidation in the retail media space is worth watching carefully. When a large holding company builds a centralized retail media practice, the question is not whether they have smart people in the room. The question is whether their financial relationships with the networks compromise the advice they give clients. Understand the incentive structure of whoever is recommending your network mix.
We have seen the same dynamic play out in programmatic media buying, and we wrote directly about how your agency’s programmatic partnerships affect media ROI. The retail media version of this problem is arriving fast.
What Good Retail Media Measurement Actually Looks Like
Good measurement in retail media is not a single metric. It is a stack.
At the top of the stack is incrementality, tested regularly and verified by a party that does not benefit from the result. Below that is category share movement at the retailer level, tracked over 12 to 26 week windows. Below that is household penetration data, which shows whether you are reaching new buyers or re-buying the same loyalists at a premium.
ROAS sits at the bottom of this stack. It is useful for operational optimization at the campaign level. It is not useful for strategic budget decisions across networks.
Brands that invert this stack and lead with ROAS end up in a cycle of over-investing in networks that look efficient on paper but are mostly harvesting demand your brand already created.
—
FAQ
What is a retail media network?
A retail media network is an advertising platform operated by a retailer that allows brands to purchase ad placements within the retailer’s owned digital properties, including their website, app, and sometimes off-site extensions. Examples include Amazon Advertising, Walmart Connect, Kroger Precision Marketing, and Target’s Roundel. These networks use first-party shopper data to target ads and report performance.
How should CPG brands measure retail media performance beyond ROAS?
CPG brands should prioritize incrementality testing, which measures sales that would not have occurred without the advertising, over last-click ROAS. Additional meaningful metrics include category share change at the retailer, household penetration growth, and new-to-brand buyer rates. ROAS is a useful campaign-level signal but a poor strategic decision-making tool across networks.
Why is incrementality testing important in retail media?
Incrementality testing separates advertising-driven sales from sales that would have happened organically. Without it, brands risk paying a premium to claim credit for purchases that were already going to occur, effectively subsidizing their own existing demand rather than growing it.
How many retail media networks should a CPG brand activate on?
There is no universal number, but most mid-sized CPG brands perform better with concentrated investment in two to three high-priority networks than with thin budgets spread across six or more. Network selection should be driven by the retailer’s share of your category’s sales volume, data transparency, and the availability of third-party verified measurement, not by relationship management or retailer pressure.
—
If you are spending on retail media networks and are not confident you know which ones are actually growing your brand, that is a solvable problem. See how we approach performance media for CPG brands or get in touch to talk through your current network mix.
—
Chris Pyne, Founder and CEO of Junction 37 and its sister venture Series A. He built Cortex, J37’s proprietary AI-driven planning ecosystem, and pioneered the integration of predictive marketing science into client strategy. Previously, Chris held C-suite roles at OMD USA and MediaCom, where he led planning for $7B in billings and 700+ employees.