Prohaska Group, the affiliate and partnership consultancy led by industry veteran Matt Prohaska, has released new research arguing that structural biases built into marketing mix modeling (MMM) are systematically penalizing affiliate marketing spend. The claim, distributed via PR Newswire, lands at a moment when brands are rebuilding their measurement stacks around MMM as third-party cookies and last-click attribution lose favor.

The core of the argument is architectural, not anecdotal. Affiliate marketing's defining commercial feature — partners are paid on performance, typically under CPA, CPL, rev-share or hybrid terms, with no media cost unless a conversion occurs — sits awkwardly inside models built to allocate incremental lift across paid channels with fixed budgets. A channel whose cost variable moves with the output variable is easy for a standard MMM to misread.

That misreading matters for program economics. When an MMM undercredits affiliate's contribution, the model's recommended budget allocation shifts dollars toward channels the model can measure more cleanly, such as paid search or linear TV. For affiliate managers running rev-share or hybrid deals, the downstream effect is real: partner commission budgets get squeezed not because partners underperform, but because the measurement layer cannot represent how they perform.

The research positions this as a measurement design problem rather than a channel performance problem. That distinction should shape how performance marketers read it. An MMM that treats affiliate as a single undifferentiated line item — rather than decomposing it by partner type, vertical and deal structure — will produce allocation outputs that conflict with what program-level EPC and conversion data already show. Any brand reconciling MMM output against affiliate network reporting should expect divergence and interrogate which side is wrong, and why.

It is worth stating plainly what this is: vendor-sponsored research from a consultancy whose business is affiliate practice. PR Newswire is a paid distribution channel. Prohaska has a clear commercial interest in brands investing more confidently in affiliate, and the release should be read as an argument to be tested, not a neutral finding. The publication did not include, in the material available, the sample size of the modeling exercises, the specific bias corrections proposed, or named brand case studies with before-and-after allocation figures. Those details would determine whether the critique holds up against brands' own triangulated data.

Even with that caveat, the underlying tension is genuine and well known to practitioners. MMM's resurgence — driven by privacy regulation, signal loss and marketer fatigue with platform self-attribution — has reopened old questions about how performance channels are valued. Affiliate is not the only channel that complains of MMM blind spots; branded search and CRM face similar critiques. But affiliate's performance-based cost structure makes it a distinctive case, and deal structures amplify the stakes. A CPL program and a rev-share program can look identical inside a coarse model while carrying entirely different margin profiles for the brand.

For networks and platforms, the research implies a workload: partner-level and deal-type data must flow into clients' modeling processes in a format MMM practitioners can actually use. For affiliates and agencies, it signals that the measurement fight has moved upstream — from last-click dashboards to econometric models that CMOs increasingly trust for budget decisions.

Prohaska says the research includes recommendations for correcting the biases; affiliates and brand measurement teams alike will want to pressure-test those corrections against their own program data before reallocating a dollar.

The full findings are expected to circulate through industry channels in the coming weeks, and the burden now shifts to brands to replicate the analysis on their own mix models.