The challenge for many companies during the Sales and Operations Planning (S&OP) process is the construction of the demand plan. This is the result of the initial stages of the monthly S&OP cycle and usually involves some activities, such as the use of statistical models to project sales from the past to the future, the adjustment of the numbers by an analyst, the review by the commercial area and a consensus meeting between various areas of the company in search of a number that is supported by all. All these activities consume time and resources, but the return is not always the desired one. To facilitate the identification of the steps that add value to the demand plan construction process, many companies have adopted an indicator known as Forecast Value Added or simply FVA.
The FVA uses a simple calculation method to determine how process steps and participants influence forecast accuracy and is used in conjunction with traditional forecast accuracy measurement indicators such as MAPE, MPE, and MAD. Figure 1 shows an example of the FVA calculation.
Figure 1 – Example of FVA calculation
Source: ILOS
Suppose a company sold 1.000 units of a product last month and, at the end of the period, it went to measure how close its prediction came to reality. In our example, the forecasting process starts with a naïve forecast, which is nothing more than establishing a simple number to calculate and that requires minimal effort, such as repeating the previous month's sales number or replicating demand. the same period in the previous year. In our example, this number was 800 and generated a MAPE of 20%. As it is the first activity of the process, there is no FVA linked to it.
Then, a computational tool, capable of testing several statistical models in the company's historical sales bases and identifying the one with greater precision, returned a forecast of 910 units, closer to what actually happened. This step, therefore, added to the process and the result is an FVA of 11%, which is the difference between the MAPE of step 1 and that of step 2. As MAPE measures error, the smaller its value, the better , and therefore the FVA in this case is positive.
After adjusting the analyst, who examined the forecast generated by the tool in search of inconsistencies in the historical sales base and managed to reduce the error by 1%, the sales area revised the numbers. As this is the area of the company closest to the customer and capable of adjusting demand to available resources, it is essential that the sales representative review forecasts considering their knowledge of promotional plans, product launches and competitor actions. In our example, the commercial significantly raised the forecast, contributing negatively to the process.
In companies with a sales team that is less engaged with the process, it is not uncommon for a revision to lead to an increase in forecast numbers, either because of the optimism characteristic of the area or because of a fear that the operations area will make a cut in the plan and do not fulfill 100% all orders. To avoid situations like this, an alternative is to link the FVA to a small portion of the commercial leaders' bonus. With this artifice, the expectation is that the commercial team, in addition to chasing the sales target, will also be concerned with the accuracy of the sales plan, encouraged by obtaining a higher remuneration.
The calculation and presentation of FVA indicators in the S&OP process make it possible to identify which steps are adding to the process and to what extent. In the previous example, although the analyst improved the forecast number, his FVA was too small. Suppose Figure 2 illustrates this analyst's monthly contribution over the past year. In some months, it is possible to observe a positive FVA, while in others the contribution was null or harmful. Is this analyst's review of the numbers more valuable to the company than the cost of his salary?
Figure 2 - Illustrative FVA monthly evolution
Source: ILOS
This is an example that a positive FVA does not always mean that the step must be maintained in the process, as sometimes the effort spent on the activity is greater than its benefit. In situations like this, it is fundamental for the company to calculate whether it is worth more to allocate the analyst to another activity in which he can add more value or to keep him in the activity to obtain an increase in forecast accuracy, even if minimal, on account of the cost of error associated with your planning process.
Due to this quest to identify waste and increase the productivity of the sales forecasting process, the FVA is considered a “Lean” indicator. Through constant monitoring of this indicator, it is possible to improve forecast accuracy without making any additional investment. Sounds good, doesn't it?
References
<https://www.sas.com/content/dam/SAS/en_us/doc/whitepaper1/forecast-value-added-analysis-106186.pdf>
<http://analytics-magazine.org/value-added-analysis-business-forecasting-effectiveness/>
Fernanda Monteiro
More than 11 years of experience in training and consultancy projects, focusing on Logistics and Supply Chain. In consultancy, he carried out projects such as Transformational Logistics Plan, Diagnosis of logistics operations, Strategy and Calendarization of Transport Operations, Measuring the Cost of Serving, Market Study, Mapping of Inventory Reduction Opportunities, Review of the S&OP Process, Management Plan Training and Implementation of Commercial Processes in companies such as Nestlé, Raia Drogasil, Ipiranga, Lojas Americanas, B2W, Coca-Cola, Andina, Embraco, Martins Atacado, Loja do Mecânico, Santo Antônio Energia, Ecoporto and Silimed. She is currently one of the teachers of the Inventory Management Course taught every six months by ILOS. She worked on the development and management of Online Courses in Logistics and Supply Chain, Supply Processes, Demand Planning, Inventory Management and Industrial Management. Still in the training area, she was responsible for applying ILOS business games in companies such as Raia Drogasil, Fibria, NEC, Novartis and Moove.
