Evolving better concepts - How “Survival of the fittest� techniques can be combined with a choice-based consumer interface to deliver better concept scores

Kevin D. Karty and Erin Breland
Affinnova Inc., United States

Michelle Gansle and Gemma Tyson
Mars Foods, Netherlands


Current innovation processes in consumer package goods and in other sectors often employ a stage gate process to manage risk in product introductions. The final stage of this process (the final “go/no-go” decision) typically relies on a volumetric projection that is based on a simulated test market (STM) tool. These tools have their roots in various models dating from the 1960s and 1970s (Assessor, Novaction, BASES, and others).

Years of data collection produced datasets that support the accuracy of sales projections resulting from these STM models. For example, multiple providers of STM models claim prediction accuracy of about +/-10% on 85% to 90% of studies. These late stage volumetric models typically incorporate product usage testing as well as actual marketing launch plans and category specific data, the combination of which substantially improves these accuracy claims. Although the inclusion of the final marketing plan in the validation cases may result in some endogeneity between the actual sales and the “projected” sales (since early data suggesting the product is not meeting expectations may cause retailers to cut distribution and marketers to shift ad and promotion spend to more promising initiatives), the accuracy claims are nonetheless substantial.