Email
sinha.rk.dr@gmail.com
At what level of granularity should you predict demand?
We know that consumer demand patterns evolve at the most granular level. The same product may witness a highly accelerating demand in a particular geography and a degrowth in another geography. The patterns evolve and change continuously.
How well do we capture it in our current forecasting process? Most companies aggregate the demand data on the product and geography dimensions for the forecasting process. Forecasts done at such high levels are then disaggregated for day-to-day supply execution.
This process evolved when collecting granular demand data was tedious and computing power was low. Monthly forecasting at an aggregate level was perhaps the only option.
Times have changed. We now have access to daily customer demand at the most granular level and the computers can process this large data efficiently to decipher the evolving patterns at that level of granularity. Should we still continue with the archaic process of monthly demand forecasting at an aggregate level?
We are leaving a lot of value on the table. Demand prediction at the most granular level is the way to go!