LogConsult Academy · I4
S&OP and Demand Planning
One plan for sales, operations, and finance, every month.
Lessons are recorded in Portuguese with Portuguese subtitles. Every lesson comes with its full text in English.
about US$ 420 · Prices in Brazilian reais (BRL), paid through Mercado Pago (Pix, credit card in up to 12 installments, or boleto). The checkout page is in Portuguese.
Program
5 video modules, 18 min, 5 exercises
- Demand forecasting and accuracy
- The monthly S&OP cycle
- Logistics capacity plan
- Crop and price scenarios
- The executive S&OP meeting
What you build
S&OP calendar and agenda for your company
Free sample: the first 2 minutes of module 1
Lesson text in English · 1
Demand forecasting and accuracy
Demand forecasting and accuracy. Hi, I'm Lia, the virtual presenter of LogConsult Academy. This is the first module of the course on Sales and Operations Planning (S&OP) and Demand Planning. Here you will learn how to build a reliable demand history, how to choose simple forecasting methods, and how to measure accuracy every month, separating random error from error that always repeats in the same direction.
Why start with demand. All of a company's planning starts from a demand forecast. It defines how much to buy, how much to produce, how many trucks to contract, and how much warehouse space to reserve. When the forecast is too high, inventory sits idle. When it is too low, product runs short and urgent freight gets expensive. That is why the first step of S&OP is to have a demand number that the whole company uses.
Sales is not the same as demand. Sales history seems like the natural starting point, but it hides problems. When product was out of stock, the recorded sale fell below what the customer wanted to buy. When there was a promotion or a very large one-off order, the sale came in above normal. Before forecasting, adjust the history to reflect true demand, marking each adjustment and the reason, so that anyone can redo the calculation.
Simple methods first. Start with simple methods that the team can understand and explain. For stable items, a moving average of the last few months is enough. For items with seasonality, such as those tied to the harvest or the weather, apply a seasonal index calculated from two or three years of history. After that, the sales area can adjust the number for launches or new customers, but this adjustment is recorded separately from the statistical forecast.
The right level of detail. Forecasting each item, for each customer, for each week, generates huge errors and a lot of work. S&OP works with aggregated numbers, such as product family and region, month by month. The more aggregated the forecast, the smaller the error tends to be, because the variations of one item offset those of another. Item-level detail is left for the planning of the coming weeks.
Measuring accuracy and bias. To improve the forecast, you need to measure the error every month. The most used indicator is the mean absolute percentage error, which shows the size of the error regardless of direction. The second is bias, which shows whether the forecast always misses to the same side. A constant positive bias usually indicates commercial optimism. Measure both with the forecast frozen at the start of the month, and not with the version adjusted afterward.
An example. Now let's look at a practical example, always with illustrative numbers. A product family had a forecast of ten thousand cases per month for three months. Actual sales were eight thousand five hundred, nine thousand, and eight thousand eight hundred cases. The average error, calculated on actuals, came to around fourteen percent. The point that stands out is the bias, because in all three months the forecast missed high.
Module exercise. Now it's your turn to apply what we covered in this module. Choose three product families from your company and gather twelve months of sales history. Adjust this history for stockouts and promotions, noting each change. Then compare it with the forecasts made at the time and calculate the average error and the bias of each family. In the next module, we will build the monthly S&OP cycle. See you there.
Lessons are presented by Lia, a virtual presenter generated by AI. The content is written and reviewed by LogConsult.