In this article, the second of the series analyzing the impact of COVID – 19 on demand planning (find here the first one), you will find a list of actions demand planners can do to correct and clean COVID’s impact on demand.
One of the first things to carry out is to identify all the items where COVID-19 has added noise to the demand history and adjust to normal levels to provide a more accurate statistical forecast calculation. The question seems clear, how do we identify this noise? First, we recommend creating work by exception alerts allowing demand planners to focus on the analysis and application of measures to face demand fluctuation. Apart from this internal analysis, external information can also help us to identify possible sources of deviations, for example, which countries have stopped their activity for a certain time due to COVID-19 irruption.
The next step would be to create or use existing product groupings such as ABC classifications to help identify and analyze the products with the highest economic impact on the company. For ABC classes, group A items represent the top 20% of the company products and usually create 80% of the revenue. When obtaining the classifications, you can also consider the option of a portfolio restructure. Reviewing group C items (usually the bottom 50% of the portfolio and only generates 5% of revenue); perhaps it is time to discontinue products with little impact on the company. Another ratio that might be useful that compares Pre-COVID and COVID demand and has a simple implementation is the coefficient of variation, CoV. This ratio will help you identify those products for which the demand does not experience fluctuations (CoV = 0 – 0.5), those for which the demand has suffered small variations (CoV = 0.5 – 1) and, finally, those for which the demand has experienced great fluctuations (CoV > 1). Comparing these results between different time periods can help you identify items that have been impacted.
Once fluctuating products have been identified, demand planners need to correct them to avoid non-realistic statistical forecasts. It requires corrections from demand planners, not only for the short term but also impacting on the long term. At this point, software such as QAD DynaSys DSCP can help them speed up the repetitive task where correction criteria could be unified. For example, QAD Dynasys has 2 popular methods to assist in cleaning and adjusting history. Either by using outlier periods by the average demand during a certain number of periods prior to COVID-19 irruption. This would smooth out the demand curve by harmonizing the signal and avoiding an irregular forecast with a low precision ratio. The other popular method would be adjusting the level of confidence by reducing the histories deviation corridor. This means the software would signal a lower tolerance to fluctuations in history. In this way; the software could then automatically adjust the demand according to previous demand planners’ indications.
After identifying history corrections, a recurring question for demand planners is at what level the adjustments should be made. Our experts believe that they should be made at the level at which the forecast is being calculated, that is if we are calculating the forecast at the site – the customer – item level this should contain all the adjustments. Although companies can ask their software providers for tools to carry out these adjustments at aggregate levels to speed up and optimize the time allocated to this task by users.
Furthermore, bear in mind that deviations in demand not only have an impact on the forecast generation but other indicators that are closely related to the company’s stock levels. The greater the demand variability and the increased uncertainty of meeting demand. The calculated minimum stock will also increase to avoid unnecessary stock-outs. In this way, it can send an increased minimum stock ripple effect down the supply chain, adjusting purchases, production, and /or distribution. Lastly and no less important, QAD DynaSys believes that identifying and differentiating all the actions carried out in the demand due to the situation of COVID-19 should be stored and subsequently analyzed in case the situation repeats itself It would be a mistake not considering a second wave of infection which would lead to another lockdown. This information can also help you to improve or create contingency plans in situations of demand uncertainty. In addition, a good analysis can provide a company detailed information about their supply chain behavior in a stressful environment and decide if any changes to the supply network or updates to suppliers or factories need to be made.
Don’t miss next week’s final post in this COVID-19 planning series, Demand Planning Against COVID-19 – Best Practices.