Structural Biases in MMM are Penalising Affiliate Marketing
New research reveals measurement flaws causing MMM to systematically misread affiliate performance—costing brands growth, efficiency, and market share.
Topics
Marketing teams have spent years refining how they measure performance, yet some of the most effective channels are still misunderstood.
Affiliate marketing sits in that gap—consistently delivering results but often undervalued when budgets are set. The issue isn’t performance, but visibility.
As measurement frameworks evolve, cracks are beginning to show, raising a bigger question: Are brands making decisions based on incomplete data, and what is that really costing them?
A new research report commissioned by Rakuten Rewards and conducted by Prohaska Consulting identifies both a critical gap in how affiliate marketing is evaluated and the clear opportunity for brands willing to close it.
The report, The Next Frontier of Measurement: Fair Evaluation of Affiliates in Marketing Mix Models, reveals that marketers have a challenging time to accurately rely on standard Marketing Mix Modelling (MMM) to assess affiliate performance.
A Channel That Doesn’t Fit the Model
Affiliates are often a proven channel, but one that doesn’t fit cleanly into the existing MMM evaluation framework. The report from Prohaska Consulting draws on more than two dozen interviews with senior marketing leaders from major brands, alongside MMM providers.
It includes analysis of real-world case studies and performance data from major retailers, agencies and Rakuten Rewards.
The central finding concludes that while affiliate marketing delivers strong returns, the models used to measure it simply aren’t capturing them.
For marketers, that may mean systematically underfunding a proven channel that reaches high-intent consumers at precisely the moment they’re ready to buy, an audience that the research identifies as difficult to reach anywhere else.
“Most marketers don’t realise how much affiliate performance is being underrepresented in their models, and that gap has real dollar impact,” said Ameet Shah, Partner at Prohaska Consulting.
“The capability to more accurately measure exists, but it’s not being applied effectively. With consistent data inputs and alignment across the ecosystem, it can be.”
The Performance Gap MMM Keeps Missing
Affiliate marketing is a proven and cost-efficient performance channel in digital advertising.
More than 80% of marketers rely on affiliates, and data from Rakuten Rewards across thousands of brands shows a 17x return in sales versus commissions. Yet affiliate marketing remains consistently misrepresented by MMM when it comes to budget allocation decisions.
The case studies in the report illustrate the cost of that mismeasurement across a range of real-world scenarios.
For example, when two brands under the same parent company ran programs concurrently on Rakuten Rewards, the actively managed program using elevated Cash Back tiers and time-limited promotions delivered up to 25x higher ROAS than its passive counterpart.
In another case, a leading agency that properly segmented affiliate inputs in MMM and planning tools experienced a 10% lift in iROAS investments on loyalty publishers that would otherwise be lost when viewing affiliate marketing as a single bucket.
The consequences of acting on incomplete information extend beyond measurement. A major retailer that paused its affiliate program after questioning its incrementality drove customers to direct competitors and lost more than 50% of its prior volume. Even a year after reactivating its affiliate program, the retailer did not recover its customers or market position.
Two Structural Flaws That Cause MMM to Misread Affiliates
Despite strong performance, affiliate marketing is often misinterpreted within MMM frameworks due to structural limitations in how the channel is measured. These gaps stem from both data constraints and modelling assumptions, which together create an incomplete picture of affiliate contribution.
Here are the two structural flaws:
- Lack of standardised data.
MMM and marketers often lump cash back, coupon, content, comparison shopping, influencer, and loyalty into one broad affiliate classification, erasing the ability to observe real performance differences.
And because affiliates are focused on conversion, they lack the data inputs that MMM relies on to measure other channels, such as impressions and clicks.
- MMM can’t do it alone.
Because affiliates earn commissions on completed sales, their activity always tracks with revenue. MMM reads that as affiliates claiming credit for existing demand rather than driving new demand.
Traditional MMM frameworks also prioritise the cheapest returns, overlooking the strategic value of scale-on-demand, one of the affiliate channel’s major advantages.
Closing the Measurement Gap
Addressing these gaps requires coordinated action across the ecosystem.
Marketers need to move beyond treating affiliates as a single channel input and instead adopt more granular measurement approaches.
Better integration of testing frameworks, richer data inputs, and alignment between brands, publishers, and measurement providers can significantly improve accuracy. The report outlines specific steps for each part of the ecosystem:
Steps for Marketers:
- Break affiliates into sub-categories within the MMM datasets rather than treating the channel as a single input
- Capture reward intensity and promotional offer changes as time-series variables (e.g. Cash Back rates, promotions, etc.)
- Calibrate MMM with sustainable testing (geo holdouts, time-based cashback tests, audience-level suppression)
Steps for Publishers:
- Provide impression and click data to enable comparisons within MMM modelling
- Build testing capabilities that support geo-targeting, holdouts, and A/B experiments
- Develop direct data integrations with major MMM vendors
The recommendation for the industry:
- Establish shared definitions and taxonomies for affiliate sub-categories, giving the channel the same measurement footing as video, native, and display
“We’ve sat across the table from brands that pulled back on affiliates because their MMM told them to do so, and they spent the next year trying to win back ground they didn’t need to lose,” said Carl Lurie Kalapesi, Chief Commercial Officer at Rakuten Rewards.
“Affiliates deliver consistently and at scale, and this research gives marketers evidence and the framework to prove it. Getting the measurement right is more than just a technical fix. It is often the difference between growing market share or handing it to a competitor.”
ALSO READ: The Monk’s Mindset: Rethinking Marketing, Trust, and Leadership
