emerging brand scoring

Abstract

Decision makers in consumer-packaged goods supply chains must decide which emerging brands to support or invest in. These decisions are often made under conditions of limited data and uncertain growth trajectories. Many prior studies focus on demand forecasting, despite evidence that brand performance is also influenced by factors such as distribution coverage, promotional efficiency, and category context. This study develops a multi-criteria decision analysis (MCDA) framework to evaluate and rank emerging brands. The framework uses demand-based features combined into a single opportunity score, with room for expansion to include additional retail performance indicators. To benchmark the MCDA rankings, an ordinal logistic regression model is used to assign brands to ordered opportunity tiers. Model performance is evaluated using a hand-labeled truth set and tested across forecast horizons of 6 to 18 months. Results indicate that median year-over-year case sales growth distinguishes low-performing from viable brands with up to 85% balanced accuracy, and that performance is stable up to 18 months in advance. The ordinal model provides finer-grained tier assignments with 93% adjacent accuracy. These findings suggest that a small set of demand-based features can support practical brand evaluation for investment and acquisition decisions.

Read the paper

View the slide deck

Forthcoming in the SMU Data Science Review.

Skills

R · Statistical modeling · Multi-criteria decision analysis · Ordinal logistic regression · Large-scale data analysis · Stakeholder communication