Marketing Mix Modeling for
CPG
%
Increase in ROI*
Increase in Sales*
Download the Industry's Toolkit
Toolkit includes: Data Request File, Marketing Mix Modeling Workflow, Industry Use Case & the Model Validation White Paper
Industry Overview
Challenges in Marketing
- Balancing short-term sales gains with long-term brand building and customer loyalty
- Attributing the effects of diverse marketing activities is challenging due to omnichannel customer engagement
- Measuring the impact of promotional activity on sales
KPIs
- Sales (by customer segment (demographics) OR by SKU OR by sales channel)
- sales volume
Customized Solutions
Promotions’ Optimization, Pricing Analytics, Brand Health Measurement, Sales forecasting, Channel ROI Measurement, Marketing Budget Allocation, Marketing Data Management, In-house Modeling Software, MMM Training, MMM In-housing Program, MMM Consultancy, seasonality impact, distribution impact, campaign level results
Client Review
Frequently Asked Questions
Q: How often do you update Marketing Mix Models?
A: For the FMCG industry, we advise on an annual or biannual update of MMMs. This is a strategic decision driven by the fact that model coefficients don’t change that often in the course of a year.
Q: How are promotions (displays/features/discounts) modeled into geographic models?
A: We use MassTer, our specialized Marketing Mix Modeling software, to build separate models for the different regions. Promotion data is broken down by region and will be input as one of the variables in the model. For different promotional tactics, they would be separated into individual variables.
Q: Weather has a massive impact on sales of a product, is it possible to include the weather factor in the model?
A: Yes, we include any external factors you deem as significant to the model including weather. We also run a deep analysis of your market to identify any other external factors that could have an impact on revenue.
Q: If I have granular regional data (by state for example), can I build different models?
A: Yes, for this, we have built a dedicated feature in MassTer to be able to use regional data to observe results on the total national level as well as on the regional level. This feature is Pooled Regression.
Q: When calculating ROIs of different media campaigns, are you looking into campaigns that are ongoing or those that happened in the past?
A: We can look into both, campaigns that have finished and those that are still running. We use historical data to measure their impacts. For those that are still running, we can account for the tail effect to forecast results. Forecasting helps you anticipate the campaigns’ ultimate impact, facilitating timely adjustments if necessary and providing a clearer picture of their potential success. (We don’t present results for the ongoing campaigns)
Q: Can we optimize our marketing spends across different products ?
A: Absolutely, we typically model each product separately to account for its unique attributes in terms of price, promotions, adstock, and saturation effects. Subsequently, we aggregate different models built using our technology, M.Insight, to optimize the marketing mix at both brand and product level.
Our Clients
