From $6,500 Risk to $30,000 in Q4: How StewMac’s AI Forecast Delivered 2.8 ROAS

The Challenge

StewMac faced a $6,500 decision: invest in a Rakuten integration representing approximately half their typical monthly affiliate budget, with no certainty of return, or walk away from an untested opportunity. Standard industry ROI calculators could not account for the brand's unique business patterns or seasonal dynamics, leaving the decision without a credible analytical foundation. The risk was real and the evidence to support the investment did not exist.

The Strategy

PartnerCentric's Business Intelligence team built a custom AI forecasting model designed to answer the specific questions StewMac's leadership needed to make the decision with confidence. The model addressed when the brand would break even, what revenue lift could be expected under conservative and optimistic scenarios, and what the true ROI of the investment would be. Rather than generic benchmarks, the model was built from StewMac's own historical data and business patterns, giving the forecast genuine specificity.

The Work

The counterfactual modelling approach predicted what would have happened if StewMac had not made the investment, providing a baseline against which actual results could be validated. The forecast projected a 30-day break-even under conservative growth assumptions, giving leadership the confidence to proceed. Performance was tracked against the model throughout, with the forecast serving as both a decision-making tool and a live accountability framework for the partnership.

Results

Rakuten broke even within 30 days, exactly as the model predicted, and climbed to a top five partner position by month two. The campaign generated nearly $30,000 in Q4 revenue at a 2.8 ROAS, and gave StewMac the confidence to onboard nine additional loyalty partners using the same forecasting methodology.

The Judges' Verdict

Judges praised this entry as an outstanding example of using AI to solve a real and persistent problem in affiliate marketing, specifically the inability to forecast true incremental value before committing investment. The counterfactual modelling approach and the accuracy of the 30-day break-even prediction were both cited as compelling evidence of a robust and reliable methodology. A practical, well-evidenced application of AI that changes how investment decisions are made.

"An outstanding submission that demonstrates the power of AI to de-risk strategic investments. The fact that the AI-predicted break-even of 30 days was met exactly proves the model's reliability."