Retail: Optimizing sales forecasts

Case study

Context and challenges

Our client, a world leader in mass distribution, had noticed sales forecast discrepancies for some cycles with no rational explanation. This made it question the overall relevance of the forecasting process, which revealed a major problem with the management of promotional offers.
In fact, the economic climate and aggressive competition was putting pressure on prices, which had caused an increase in the number of sales promotions.
This flow, which accounted for 20% of sales, was subject to sophisticated engineering and a variety of customer supply methods.
The Supply Chain Department commissioned Argon Consulting to diagnose these processes. Our task was to map and analyze the existing sales forecasting method and to identify potential improvements.

Approach and drivers

1. Mapping of existing processes​

To provide an accurate description of existing processes, Argon teams conducted one-on-one interviews with the various stakeholders: forecasters, sales development, customer services and marketing. These interviews were used to create an inventory of management rules and modeling criteria. The diagnosis was also based on specific stakeholder feedback.

Three main sales promotions related review cycles were mapped:

  • Yearly process: approval of promotional budget and definition of the promotional plan
  • Monthly process: joint approval of sales promotions
  • Weekly process: update of forecasts based, in particular, on customer information

The forecasts included five main components: the baseline, promos, the inventory effect, anticipation and safety inventory, and medias.
At each stage, the components of the forecast were mapped, to identify the stakeholders involved and the management rules employed. This snapshot was the working basis for listing potential improvements. Similarly, the sources were clearly identified along with the error risk related to
each of them.

2. Diagnosis of operational strengths and weaknesses​

The main asset of our client’s current process was the involvement of all stakeholders.
However, improvements still needed to be made because the roles and responsibilities of each stakeholder were not satisfactory, and the combination of inputs from different entities had its limits. The frequency of updates and the forecast approval procedure were structural elements of the process that needed improvement to become more accurate. Additionally, very little feedback was available and there were few continual improvement initiatives.

3. Identifying areas for improvement​

The diagnosis indicated drivers for improvement in five main areas:

  • Forecasting process​

Managing and implementing a peer approval process for forecasts and cross‐functional meetings, defining the inputs update frequency for different forecast horizons and assessing the impact on management

  • Management and continual improvement​

Implementing and managing key performance indicators, implementing cross‐functional feedback and capitalizing on the developments made

  • Organization

Reviewing stakeholders’ roles and responsibilities

  • Ensuring the forecast accuracy​

Assessing the benefits of more macro‐level analysis and ensuring the accuracy of all forecast components

  • Information Technology applications

Implementation of a model facilitating a statistical analysis of the number of products included in the sales promotion and development of the existing application to allow for the integration of inputs at a more detailed level

Each driver was detailed, with its related challenges and accessibility, which made it possible to prioritize actions by demonstrating “quick wins”.


The diagnosis performed by Argon made it possible to identify the strengths and weaknesses of the current process and to understand the causes of the forecasting discrepancies.
The roadmap of the actions to be taken to ensure the accuracy of the process, regarding sales promotions in particular, was shared and approved by all stakeholders.