Business Context
A leading consumer packaged goods (CPG) company was facing challenges with
their ongoing trade and consumer promotions, which often failed to deliver the
expected incremental sales and profitability. The key issue was the lack of a clear
understanding of how different promotion mechanics, discount levels, and scenarios
impacted overall product portfolio performance. To address this, the company
partnered with AiQMEN to develop an optimized promotions strategy using
advanced analytics, aimed at maximizing both sales and profit across their portfolio.
AiQMEN’s Solution
The project scope initially covered two key product categories across two markets.
AiQMEN implemented the solution on Microsoft Azure, utilizing the platform's
scalability and seamless data integration capabilities. Key data sources included
internal shipment data, syndicated sales data (Nielsen), the promotions calendar,
financial data, product hierarchy, as well as external factors such as weather data
and local events.
The data was harmonized to ensure consistency across these diverse sources,
aligning everything at the SKU x week x market x retailer level. After harmonization,
the data was transformed to facilitate the development of an advanced analytics
solution. Through exploratory data analysis (EDA), key trends and patterns were
uncovered, which informed the development of initial hypotheses.
To evaluate the effectiveness of promotions, AiQMEN built mixed-effects regression
models for each product category and market. The various key performance
indicators (KPIs) were randomized at appropriate levels, ensuring the model was
robust. Comprehensive model validation steps were undertaken to prevent overfitting
and ensure reliability.
The models delivered valuable insights into sales volume drivers such as product
distribution, pricing, discount percentages, weather, and events. To support decision-
making, AiQMEN developed a custom web-based application, offering users a
historical performance dashboard and a simulation module for conducting what-if
scenario analyses. This allowed business users to test future promotional strategies
and evaluate their potential impact on sales and profitability.
To streamline the process end-to-end, the entire data pipeline was automated,
incorporating MLOps capabilities for monthly updates with minimal manual
intervention.
Impact
The solution enabled the client to evaluate various promotion scenarios more
effectively, resulting in refined and more impactful promotional strategies. By
identifying underperforming promotions, the company was able to discontinue those
that failed to generate positive incremental volume or profit. With these data-driven
insights, the client was also able to negotiate more effectively with retailers.
As a result, the company achieved ~ 6% increase in incremental revenue, driven by
strategic adjustments to its promotions approach.
Found interesting! Want to discuss more? Let’s talk
