Demand Planning Amid the Coronavirus Outbreak
Welcome back to Demand Planning Amid the Coronavirus Outbreak, a series by Fiddlehead Technology. The purpose of this series is to support S&OP teams across the food and beverage industry who are assuring supply chains can feed consumers around the world who continue to depend on them every day to eat.
In part two of the series, we offer insights on how S&OP teams can strategically use POS data to gain a better understanding of their short-term forecasts.
Part Two: Transitioning to Strategic Forecasting
Nearly a month into the coronavirus outbreak across North America, the question on many minds is how long will this last? The virus has upended the economy as ripple effects continue to be felt across all sectors.
As seen through the lens of food and beverage manufacturing, the pandemic has caused major shifts in channel mix and purchasing behaviour, as reactionary decisions by consumers continue to amplify a bullwhip throughout supply chains.
The IBF reported, during their town hall: Predicting Consumer Demand During a Crisis, the outbreak has led to 40% of CPG companies moving planning to week-to-week cycles and 10% of CPG companies moving away from an S&OP process completely.
Demand planners are busy reconciling their forecasts, while attempting to making sense of the demand volatility, consumer behavioural changes, and addressing how to treat the current period of demand data moving forward.
The pandemic has saddled S&OP team with a daunting workload. Here are five thoughts to help you meet these challenges head on and move from a reactionary to strategic forecasting practices.
Deviate from legacy system forecast practices.
POS data is your truest consumer demand signal available and guards against false signals such as rationing and order batching.
Transition to a short-term forecast driven by POS demand history as a leading indicator. This will capture demand pattern anomalies which your legacy systems may fail to predict.
The demand shock caused by the pandemic varies greatly by product line and geographical location, necessitating a granular approach to your forecast. Set this forecast on a temporal series of one to eight-week intervals. The hierarchical series should be set to the SKU/location level.
Your ability to operate hinges on the weakest link in your supply chain, and though tedious, greater disaggregation provides heightened demand insights and operational agility.
Understand behavioural changes = utilizing POS data.
Most of you receive POS data from retailers on a daily or weekly basis. It is paramount to leverage this data in understanding the behavioural changes your customers are exhibiting.
By revising future forecasts using POS demand history as a leading indicator, you will illuminate your products’ demand patterns relating to the average market basket mix. This frees you to focus on forecasting the lower product mix as it will indicate which items have the highest demand velocity.
As a result, your company can more accurately predict supply replenishment to your retail customers.
Data for alternative demand signals
Now more than ever it is important to have visibility into your downstream supply chain, because demand is changing so rapidly. In addition to leveraging customer POS data, consider alternative sources for demand signals like syndicated consumer panel and retail sales data, foot traffic data, and mobile app usage, as leading potential indicators for your forecast, shifting consumer purchase behaviours in your categories, and early signs of recovery.
Cleansing the coronavirus impact from historical data?
Once the pandemic subsidizes and consumer behaviour reverts to typical demand patterns, the question facing many demand planners Is whether it is best to cleanse or not to cleanse the virus’s impact from historical data?
While the decision is more nuanced for each demand planner, the consensus is if you are focused on short-term averages reflect the most recent demand signal and refrain from data cleansing; while, if you are focused on intermediate and long-term ranges revert to typical seasonality and cleanse data.
Though, this decision must not be rushed as you may be erasing key information, or over/under-projecting what the normal demand would have been in absence of the outbreak. We caution against manually removing abnormal historical demand. Rather, by adding dummy variables to existing models you can capture these abnormal demand patterns and automatically optimize the historical demand to reflect normal demand patterns.
Regardless of the cleansing strategy you choose to adopt, we recommend that you keep an archive of the actuals for reference in the event there is a resurgence of this pandemic or another in the future. The lack of data for historically comparable events for the current pandemic is a big part of the current challenge that FMCG companies are facing. Keeping logs of actual sales order or shipment data, as well as your company’s tactical planning decisions during this time, in addition to your adjusted history, will provide useful future benchmarks.
Apply learnings to build more resilient demand planning.
A positive outcome from the pandemic can begin with you transforming your S&OP process to be more resilient to demand shock.
Considering the potential for a second and third waves of the pandemic, use outlier variables to mark recent demand history. Assure you have created the opportunity for future extrapolation if we face a similar situation over the coming years.
Ask yourself, beyond your S&OP process, are you sufficiently embracing technology from an operational perspective? Could your company reduce the burden of labor relating to warehousing and transportation? Use this situation to spark a broader conversation with your colleagues to create pandemic-proof supply chains across the entire food and beverage industry.
What’s Next: Stay tuned for Fiddlehead’s next post – Part Three: Alternative Indicators for Forecasting. This is the final instalment of our series and to be released Wednesday, April 29th, 2020.
About Fiddlehead Technology: Fiddlehead works with prescriptive analytics to find elegant solutions to some of the food and beverage manufacturing industry’s most complex problems. The result is more accurate demand forecasts, allowing companies to lower inventory, achieve higher levels of service, and improve their margins in an increasingly complex global market.
