Email
sinha.rk.dr@gmail.com
Should we predict demand at the granular level or at an aggregate level?
I am again stirring up a hornet’s nest. There are strong views on either side. Most practitioners of TOC and DDSC avoid predicting demand at the granular level due to its inherent high variability. On the other hand, there are practitioners who insist that demand must be predicted at the granular level for smooth running of the supply chain. How do we resolve it?
If we predict demand at a higher level, we must disaggregate it to more granular levels where inventory buffers are kept. The process of disaggregation often induces more inaccuracy than the additional variability experienced at the granular level.
In my opinion, we should predict demand where the front line inventory buffers are kept for demand fulfilment. This would avoid the need for disaggregation. Would we lose out on accuracy? Not really, if we modify the process of demand prediction.
If you aren’t convinced, I would like to rest my case with a live example that all of us can relate to, which clearly shows how predicting at the granular level improves the prediction accuracy significantly!
Wait for my next post…