Demand Forecasting or Demand Planning is a critical organization process that ties together the commercial and manufacturing/delivery sides of the organization to ensures that customer’s needs are met with the right product at the right time as well as the company is running effectively. An accurate and effective demand forecasting/planning function not only provides stability to sales and revenue projections, but it also ensures that inventory and warehousing management run optimally and that customers’ needs are satisfactorily met.

Typically demand forecasts are based on analysis of sales in previous time periods using relatively straightforward statistical and time-series techniques. There is also a need to adjust these forecasts for the impact of seasonal trends and events, competitor promotions, economic conditions, etc – which are cyclical, predictable, and thus possible to model for based on historical data.

Like most other aspects of life and business, the COVID-19 situation has thrown otherwise predictable demand patterns and forecasts completely off track. Whether you belong to an industry that has been impacted positively (think food and beverages) or one that has suffered sluggish demand, chances are that you do not know how to play your inventory, manufacturing plans, shipping schedules, or workforce planning for the coming 1 to 3 months. You are not alone, most demand planners and have not seen such ‘black swan’ / VUCA events in their lifetimes and do not know where to start.

Complex machine learning algorithms that are robust and reliable in ‘normal’ times, will not prove so useful now given their reliance on historical data and trends, which for this situation does not exist. Now is the time to put good old human intervention back to work.

Given the speed and agility with which decisions need to be made and changes need to be effectuated, companies need to adopt simple and pragmatic approaches. In our experience, Scenario Forecasting is one such approach:

STEP I: Look back in data (both external and internal) to find extreme events, positive and negative ( like MERS, SARS) to assess the peak and troughs of demand

STEP II: Create scenarios, perhaps even with multiplier effects if the past event does not justify the extremity of the current situation

STEP III: Create an integrated demand forecast with different scenarios and planning of the areas in the vicinity of the one likeliest to materialize. While this is not a perfect solution, operating within a bandwidth (based on data) allow for a certain degree of planning (compare to abandoning the forecast altogether and relying on pure intuition)


Even after this situation has passed, organizations will most probably not be able to return to their old ways and models. They will have to readjust their models (not only in terms of unlearning from data of this period) but also adjusting to the shift in the behavior of markets and consumers.

So brace yourselves for the long and curvy road ahead!