How a leading global CPG brand linked its online and offline revenue models to trace retail media’s reach, and found that one retailer’s media also lifts sales at another.
- At a glanceSector
- Consumer packaged goods (CPG)
- Scope
- One global CPG brand, standard media plus retail media across multiple online retailers
- Engagement
- Ongoing
Always-ON measurement, delivered on the MassTer platform.
The challenge
For a CPG brand selling through standard media (social, TV, out of home) alongside retailer-run media, including Amazon-style retail media, measuring what actually drives revenue is harder than it looks. Retail media sits at the boundary of two blind spots. Attribution methods measure one sales channel at a time, so a retail media campaign’s pull on brick and mortar sales goes uncounted. And any lift one retailer’s media gives to a different retailer’s sales, a halo effect, is invisible to a channel by channel view.
Attribution methods measure a single sales channel at a time. That breaks down for a brand running retail media across several online retailers while also selling through brick and mortar stores, because the two effects that matter most here, retail media’s pull on offline sales and one retailer’s media lifting another retailer’s revenue, both cross channel boundaries attribution never reaches. Left unmeasured, both effects get folded into baseline or dismissed as noise, and every media decision built on that view inherits the same blind spot.
The solution
MASS Analytics built two linked marketing mix models rather than one, a model for online revenue and a model for offline revenue, both fed by the same five data categories: earned and owned digital media, paid digital media, onsite and offsite retail media, non-media activity, and seasonality and external factors. Every online retailer the brand sells through was modeled individually inside the online revenue model, and the same retail media variables were carried into the offline revenue model rather than dropped at the channel boundary.
| Measurement lever | Model design | What it measures | Finding |
|---|---|---|---|
| Retail media’s offline reach | Onsite and offsite retail media variables retained in the offline revenue model, tested for significance | Retail media’s contribution to brick and mortar revenue, not just online revenue | Onsite and offsite retail media together drove 10.5% of total media-attributed offline revenue |
| Cross-retailer halo effect | Each retailer’s media entered as a variable in every other retailer’s revenue model, not only its own | The indirect lift one retailer’s media gives to a different retailer’s revenue | Onsite media drove up to 45% of its own retailer’s revenue, with a 12% halo lift onto a second retailer |
| Cross-channel model design | Two linked hierarchical models, online revenue and offline revenue, sharing the same media and non-media inputs | Whether a measurement approach can follow a single media dollar across both sales channels | Enabled measurement across sales channels, something the previous channel-by-channel view could not do |
Retail media and cross-retailer effects entered both models as explicit variables. Standard media, non-media activity, and seasonality stayed in the model with standard MMM treatment.
Beyond each retailer’s own media variables, the online revenue model for every retailer also carried the media variables of every other retailer the brand sells through. This let the model estimate whether a campaign run on one retailer’s platform moved revenue on a different retailer’s platform, the halo effect, rather than crediting that lift to brand awareness or treating it as unexplained residual.
Results and Impact
The linked model design produced two findings a channel by channel view could not surface. Onsite and offsite retail media together drove 10.5% of total media-attributed offline revenue, a contribution single-channel attribution would have missed, since it never measures retail media against brick and mortar sales.
10.5%
of media-attributed offline revenue, from onsite and offsite retail media
40%
of Online Retailer A’s own revenue driven by its onsite retail media
45%
of Online Retailer B’s own revenue driven by its onsite retail media
12%
halo lift onto Retailer B’s revenue from Retailer A’s onsite media
The 10.5% figure is onsite and offsite retail media’s combined share of media-attributed offline, brick and mortar revenue, measured by retaining those variables inside the offline revenue model rather than treating retail media as an online-only lever. It matters because a large share of shopping behavior already crosses that boundary: about 54% of shoppers research a product online before buying it in a physical store, according to Shopify.
Onsite retail media drove up to 45% of its own retailer’s online revenue at the high end, and the cross-retailer halo measured a 12% lift onto a second retailer’s revenue from the first retailer’s onsite media. Left unmeasured, that lift would have gone uncredited or been wrongly attributed to the retailer where the sale landed. Measured directly, it gives the brand a real basis for negotiating retail media spend across partners, not just what shows on one platform.
“We knew retail media was working somewhere. We could not have told you it was also working somewhere else.”
A MEMBER OF THE BRAND’S MEDIA ANALYTICS TEAM
Want the same view of your retail media?
MASS Analytics builds MMM models that separate every commercial lever, from paid media to retail partnerships, and refresh as often as the business changes. Talk to our team.
Book a demo
