Mitigate risks by implementing the right methods, appropriate tools, and an adapted process!

Establishing reliable forecasts for your sales and other variables related to your activities (cash flow, production, revenue, etc.) is crucial for effective business management and performance. Otherwise, for instance, insufficient stock and delivery delays can result in revenue loss and a poor customer experience.

Such issues arise when sales are not accurately forecasted and planned. Reliable, well-monitored, and widely shared forecasts within the company, facilitated by a strong collaborative approach, help guard against various risks associated with the market, competition, and your customers.

It is important to be guided and advised by experienced experts in the field of Data Science, Forecasting, and Predictive Analysis. 

A Use Case in the Agri-Food Industry

We assisted one of our clients in the agri-food industry in implementing a new sales forecasting methodology by product (SKU), segments, and product families, across various geographic areas (regions in France and some European countries), and by customer type.

We also supported them in enhancing the forecasting process by documenting it, making it more collaborative, and ensuring better control.


  1. Discrepancies between projected and actual sales have widened, leading to disruptions in the supply chain.
  2. Reaction and corrective actions are delayed due to an outdated process, and customer expectations have significantly changed.
  3. The forecasting methodology is also outdated: relying on simple extrapolations of past trends, incorporating seasonality, it no longer aligns with rapidly evolving customer trends and expectations.
  4. Forecasts were not shared among various stakeholders and were predominantly validated at the central level.

The implemented solution:

  1. Data audit conducted on ERP data, forecasting methodology, and end-to-end process.
  2. Restructuring of essential data for forecasting, including the integration of new external data such as competition data, satisfaction surveys, etc.
  3. Implementation of a new forecasting software to replace the « Excel » method. This tool offers a variety of adaptable forecasting methods for each product type. A recent product cannot be forecasted using a traditional extrapolation method.
  4. Various modeling, testing, and iterations have helped reduce discrepancies between actual and projected sales.
  5. The ‘qualitative’ perspective of the sales force has been integrated into the modeling, enhancing forecasts by incorporating on-the-ground insights.
  6. Sharing of initial and final forecasts with various stakeholders in the forecasting process has been strengthened, establishing an effective collaborative approach and method.