By Michele Grossman, Founder MRA LLC
& Wayne Viljoen, Sr. Director RGM and Analytics, MRA LLC
Introduction
Many of our CPG clients have noted that “Base Trend” is accounting for an increasing share of their sales decomposition outputs. This presents a challenge for planning and execution, because:
- Primary commercial drivers may no longer explain the bulk of sales variation, or
- Some of these drivers are inadvertently captured within the ‘Base Trend’, obscuring their true impact, or
- The ‘Base Trend’ itself may need to be further disaggregated to reveal underlying dynamics – both positive and negative.
This paper outlines Market Research Alliance’s (MRA’s) proprietary approach to resolving this issue and improving the clarity, precision, and actionability of sales decompositions.
Table of Contents
1. A growing challenge in sales decomposition
2. A Robust approach to understanding sales variation:
2.1 Key Driver Analysis
2.2 Sales Variance Decomposition
3. What might be hiding in the ‘Base Trend’?
4. Partner with MRA
1. A growing challenge in sales decomposition
Sales decomposition outputs often offer a helpful high-level view but lack the granularity needed to guide strategy. A growing share of volume change is attributed to an ambiguous ‘Base Trend”, limiting visibility into the true drivers of performance.
While this level of simplicity may have worked in the past, by accurately accounting for some of the key commercial levers, today’s market complexity demands a more robust and actionable approach.
2. A robust approach to understanding sales variation
MRA conducts sales variation decomposition in two ways:
2.1 Key Driver Analysis
A statistical approach to uncover why unit, volume, or $ sales changed:
- Coefficient-based approach using Machine Learning (ML) to reflect the relationship between changes in volume due to changes in key drivers
- Flexible approach where the drivers are customized to reflect the ones that most account for sales variation
- This model can incorporate different data sources into a single due-to
2.2 Sales Variance Decomposition
A mathematical approach to quantify how much dollar sales changed due to fixed commercial levers:
- Decomposes $ sales variance into price, distribution, velocity, and mix effects
- Split into non-promoted and promoted buckets
- Accounts for 100% of sales variation (i.e., no ‘Base Trend/ Other Drivers’ bucket) – but velocity contains the effects of different drivers (as seen in A. Key Driver Analysis)
While ‘B. Sales Variance Decomposition’ provides insight into the health of the growth algorithm of a given category, segment, manufacturer, or brand, ‘A. Key Driver Analysis’ provides a diagnosis into the effect of a range of business drivers. Depending on the business objective, A, B, or both can be used.
Due to the proprietary approach of MRA, POS data can be supplemented with additional data – from other data sources – to generate a comprehensive breakdown of sales. This approach accurately reflects the ‘Base Trend’ impact while splitting out the effects of other variables.
3. What might be hiding in the ‘Base Trend’
Using MRA’s method, we identify drivers often buried in “Base Trend” metrics, such as:
- Competitor pricing changes
- Competitor distribution changes, and disruptive innovation
- Changes in media and marketing spending
- Changes in shopper marketing spending
- Macro-economic conditions
- Changes in shopper behavior
- Seasonality, holidays and weather events
- Incidence rates of illness
- Out-of-stocks or transportation disruptions
- Cyber-security issues affecting supply
- Changing tariffs and other import issues
- Latest wellness trends – e.g. adoption of GLP-1’s
4. Partner with MRA
We’ve helped clients reduce the unexplained portion of volume shifts in their volumetric models, improving forecast accuracy, trade spend allocation, and supply chain efficiency.
If you’d like to dig deeper, contact us anytime. We’d be happy to discuss how we can tailor this approach to your needs.
