1. As a practice, most of the supply chain efforts were spent on Manufacturing & Distribution. In the same context, the goal of the supply chain planner was to have the distribution centers replenished with the SKU’s; the manufacturer would leave it to the retailers to pull based on their (retailers) requirements.
2. In contrast to the above, the Flowcast approach is to start planning at the store level. The thrust is to spend efforts to plan at the store level & produce so called “true” forecasts of the demands, with the end objective of reducing out-of-stock situations, improving inventory performance in terms of increased inventory turns & finally, an increased ROI from the supply chain network.
However, the aforesaid approach of Flowcast suffers from some pitfalls & the below mentioned paraphrase would attempt to convey my thoughts on the pitfalls:
1. Flowcasts tend to be wrong: Although this principle applies to Forecasts too, it is more applicable to Flowcasts since the concept of Flowcasts hovers around on generating a forecast based on the end item sale. As a general rule of forecasting, forecasts tend to be less accurate for end product items than for a product family or a group of products.
2. Though Flowcasts tend to overcome the hurdle of data accuracy since the data is pulled directly from a Point-Of-Sale (POS) system, however, the hierarchical dimension of the data does not seem to be agreeing well with concept of forecasting.
3. To exemplify, consider only the product hierarchy – instead of Flowcasting the sale of the number of units of X brand of a specific flavored cream biscuits, a forecast of the sales of the number of units of the entire cream biscuits category would be less erroneous.
4. However, the next logical question a retailer would have is the number of each brand of cream biscuits he (or she) would need to stock up at the POS.
5. This is where Flowcasts would help to check the past trend of the sales & apportion the gross figure of cream biscuits appropriately. That’s fine enough!
6. However, Flowcasting cannot be used independently. The entire theory of Flowcasting relies on the premise that the corresponding DC’s & Warehouses are the supply points whereas the specific POS is the demand point. Hence, a forecast will still have to be generated at the DC / Warehouse level to ensure smooth supply of the items when demand exists. Of course, the Flowcast data (past POS data) can serve as input for the forecasting at the DC / Warehouse level.
7. Moving on, the time horizon associated with Flowcasting will have to be very low, compared to Forecasting. Rule wise, this is in favor of Flowcasting, but considering that the related forecast at DC level will have to be also generated for a shorter time horizon.
8. Lastly, Flowcasting can be a great advantage for manufacturers who own the POS. In cases of hybrid supply chains, typically with retailers, the production numbers for the manufacturer would go for a toss, considering that the manufacturer would have to account in all the POS when generating a production plan.
9. While it can be argued that Flowcasting would inculcate the much talked about Pull strategy of replenishment, but considering the limitations of production lead times & those of production schedules, it may be highly infeasible.
10. Further, it may be argued that Flowcasting inputs can help the manufacturing schedulers to produce based on the customer demand patterns by incorporating the trend & seasonality in their production plans. Well, for these, there are various Statistical forecast engines which can forecast after taking into consideration the trend & seasonality patterns.
