• Written by Azhar Aliyar, Director

A category manager for a major FMCG brand is reviewing sales data to inform their next promotional campaign, but the numbers are not adding up due to the complexities of promotional distortion in their historical data. This is a common problem in the FMCG industry, where high volume, short shelf life, and promotion-driven demand can make it difficult to get accurate insights from data.

FMCG complexities

The FMCG industry poses unique challenges for data analysis, particularly when it comes to promotional distortion. Promotions can create artificial spikes in demand, making it difficult to determine baseline sales. Additionally, short shelf life products can lead to write-offs, which can further distort the data. Distributor visibility is also a challenge, as it can be difficult to track inventory levels and sales data across multiple distributors.

These complexities can make it difficult to develop accurate models for forecasting demand and optimizing inventory levels. For example, a model that is not accounting for promotional distortion may overestimate baseline demand, leading to overstocking and wasted product. Similarly, a model that is not accounting for shelf life write-offs may underestimate demand, leading to stockouts and lost sales.

Field-level meaning

Another challenge in FMCG data analysis is ensuring that the data is properly defined and documented. Column names are not sufficient documentation, as they may not accurately reflect the meaning of the data. For example, a column named "sales" may actually represent shipments, rather than actual sales. This can lead to incorrect assumptions and models that are not accurate.

Field-level meaning is critical in FMCG data analysis, as small errors can have significant consequences. For instance, a model that is using the wrong definition of "sales" may be optimizing for the wrong metric, leading to suboptimal decisions. It is essential to have clear and consistent definitions of each field to ensure that the data is accurate and reliable.

Systems out of scope

FMCG companies often have multiple systems and stakeholders involved in their operations, which can make it difficult to get a complete picture of the data. Some systems, such as transportation management or warehouse management systems, may be left out of scope, leading to partial answers and incomplete models. This can result in suboptimal decisions and lost opportunities.

For example, a category manager may be using data from the enterprise resource planning (ERP) system to inform their decisions, but may not have access to data from the transportation management system. This can lead to a lack of visibility into inventory levels and shipment schedules, making it difficult to optimize inventory levels and minimize stockouts. To get a complete picture of the data, it is essential to integrate data from all relevant systems, including those that may be outside of the traditional scope of data analysis.

By integrating data from all relevant systems, FMCG companies can get a more complete picture of their operations and make more informed decisions. This can be achieved by working with data integration specialists who can help to identify and integrate data from all relevant systems.

Gaps in the data

Gaps in the data, such as stockouts, backfills, and other anomalies, can also teach a model the wrong lesson. For example, a model that is trained on data that includes stockouts may learn to predict stockouts as a normal part of the business, rather than as an anomaly to be avoided. This can lead to suboptimal decisions and lost opportunities.

It is essential to identify and address gaps in the data to ensure that the models are accurate and reliable. This can be achieved by using data quality checks and data validation techniques to identify and correct errors in the data. Additionally, it is essential to have clear and consistent definitions of each field to ensure that the data is accurate and reliable.

By addressing gaps in the data and ensuring that the data is accurate and reliable, FMCG companies can develop more accurate models and make more informed decisions. This can be achieved by working with data analysis specialists who can help to identify and address gaps in the data.

Limitations of this approach

This approach to data analysis is not a silver bullet, and there are limitations to its effectiveness. For example, it may not be suitable for companies with very small datasets or those with highly complex operations. Additionally, it requires significant expertise and resources to implement and maintain.

It is essential to carefully consider these limitations before implementing this approach. Companies should carefully evaluate their data and operations to determine whether this approach is suitable for their needs. They should also consider the resources and expertise required to implement and maintain this approach, and ensure that they have the necessary capabilities in place.

By carefully considering these limitations and evaluating their data and operations, FMCG companies can determine whether this approach is suitable for their needs and make more informed decisions about their data analysis strategy.

• Companies with very small datasets may need to consider alternative approaches
• Companies with highly complex operations may need to consider more specialized solutions