How to make adjustments to demand planning in the post-pandemic

Demand planning, in Covid-19 times, suffered a huge impact. Many companies that noticed unexpected variations in their demands, sometimes upwards (sanitizing products, ready-to-eat foods), sometimes downwards (air tickets, tourism), had to turn off their algorithms and rethink the planning process, trying to “look to the future” ” and forget about historical data for a bit. At the end of the quarantine, the planning process should return to the “new normal”, and adjustments to historical data will be necessary.

Figure 1 – During the pandemic, demand planning had to be adjusted, as past sales data do not reflect the effects during the pandemic. After the crisis, adjustments will need to be made to the historical data for the normal process to return. Image: Jan Vasek by Pixabay

To deal with this unprecedented change in demand behavior, we at ILOS made a Live in early May to discuss what adjustments to the demand planning process would need to be made during and after the pandemic and the prevailing trends for the future. A discussed and fundamental point is: when should we consider returning to the traditional planning process? And what adjustments need to be made to the historical series, so that we can return to the use of forecasting algorithms?

To know when we should return to the monthly demand planning process, we suggest calculating and monitoring the daily (or weekly) coefficient of variation (CV) of the product demand curves. The coefficient of variation is the ratio between the mean standard deviation and the mean demand, so it is an indicator of the variability of the demand curve. During the pandemic, CV increased a lot, as there was great variability in demand, so it is to be expected that, with the end of the pandemic, this indicator will return to previous values. This, therefore, is an indication that the return to the original process can be performed (Figure 2).

Figure 2 – The calculation of the CV of the historical series indicates the moment to return to the use of sales forecasting algorithms. Source: ILOS

For the adjustment of the historical series, it is important to understand the components of the sales series (level, trend, cycle and seasonality) and to analyze what the future behavior of the series will be, according to the expectations and measurements of changes in the consumption pattern of certain products, categories and segments. For example, in the case of toilet paper, where there was a momentary increase in sales, but with the consumption pattern maintained, an adjustment to the seasonality of the series should be sufficient. Comparing the seasonal index of the month in previous years with the seasonal factor of March/20 is a way of defining which multiplicative factor should be applied to the data throughout the crisis. This same procedure can be applied to demands that have suffered an abrupt drop, in which a return to the previous level of sales is expected after the coronavirus crisis.

As for sanitizing products, there was also an abrupt increase in sales due to the need for constant hygiene, but a higher sales level than that seen before the pandemic is to be expected, as people will become more concerned with cleaning and hygiene. In this case, in addition to adjusting for seasonality to deal with this sales peak during the crisis, it will be necessary to correct the level of the sales series, as the consumption pattern will lead to greater use of the product, in a perennial way (Figure 3 ).

Figure 3 – Baseline adjustment for products whose demand was affected by the new coronavirus. Source: ILOS

Monitoring the coefficient of variation of the historical series to define when we should return to the traditional planning process, identifying the behavior of demand in the new reality and making the necessary adjustments will be essential to get out of the crisis with a regularized demand planning process. It's worth checking out our live stream to understand what trends are expected for demand in the future, and how to reevaluate the planning process throughout the crisis.