Advanced Sales Promotion Optimization with Logit and Semi-Log Models for Detergent Brands
Conducted a comprehensive analysis of 177 households over 135 weeks to assess how price cuts, display promotions, and feature ads influenced consumer purchasing behavior and profitability for leading detergent brands: Wisk, All, Tide, and Cheer. Key findings included a 190.35% increase in purchase likelihood from price cuts and a 59.9% rise in retailer gross profit through optimized pass-through strategies and promotions.
This project analyzed the effectiveness of various sales promotions on purchase quantities and profit margins for liquid detergent brands. The study leveraged scanner panel data from 177 households over 135 weeks, focusing on Wisk, All, Tide, and Cheer.
We built and analyzed three models using SAS:
Binomial Logit Model:
Purchase likelihood increased by 190.35% for a $1 price cut.
Display promotions boosted purchase odds by 63.17% and feature ads by 74.5%.
Multinomial Logit Model:
Tide had the highest consumer utility, outperforming other brands, while Wisk performed the lowest.
Semi-Log Models:
Display promotions increased purchase quantity by 244.54 oz. for Tide, compared to 70.42 oz. from price cuts.
Skills and Tools Utilized:
Software: SAS, Excel
Techniques: Binomial Logit Model, Multinomial Logit Model, Semi-Log Model
Analytical Skills: Statistical modeling, sales performance analysis, predictive modeling, data interpretation
Power in Numbers
Stores Analyzed
Optimal Price($)
% sales drop for a $1 price increase without promotions