Statistically Orchestrated Forecasting to Achieve Real-time Inventory Management

The problem at hand: One of the most crucial aspects affecting the working capital fluctuations and EBITDA (i.e. profitability) levels in the manufacturing industry connects to the inventory management. In general our inventory management still largely relies upon the field sales / dealership / distributors indents or demands fed into the demand management channel. The prospect theory comes to the play and the front ending teams out of fear of losing on any sales opportunity keeps a cushion while preparing a demand indent, now this cushion travels all along the chain to reach to the distribution warehouses ("DC's") to the Finished Good ("FG") warehouses to the sales and operations planning ("S&OP") team which in turn triggers the dispatch schedules and the manufacturing unit following a basic Min-Max principle.

Here the issue of inventory piling up further takes a hit on account of Full-Truck-Load fallacy ("FTL Fallacy") i.e. sending the excess material compared to what is demanded as the truck fill-rate will suffer translating into err in forecasting as the baseline is decided by the Min-Max philosophy. So we keep padding up the demand as we travel down the chain, resulting into higher Days of Inventory, extending the cash conversion cycle ("CCC") and stressing the financial management and profitability deliverables.

The above very short analysis of some of the listed major manufacturing companies highlights the days of inventory overhang of 137+ days translating into an average cash conversion cycle of 157+ days. This is the cost component which constitutes 7% of India's overall logistics costs followed by transportation. The irony of the situation is usually that even with relatively higher inventory levels we still face stock-outs and service level failures.

Root Cause of the Problem: The information latency, visibility and time lag are the primary root cause of the issue. usually even with an ERP/SAP based system the real-time data visibility remains a challenge and the teams still communicates over the emails, SMSs or whatsApp. The key question remains - the non availability of a parallel and end-to-end visible system which can be as good, as live and as recent as a WhatsApp message. The prime reason the unorganised digital communication media taking over the formal and integrated channel, however the latency is removed but at the cost of extremely isolated understanding and context.

This entirely results into falling back on the most basic and safer cushion based Min-Max inventory philosophy piling up the issue further.

LATENCY+LAG in the system => Adapting Isolated but faster communication channels viz. SMS, WhatsApp, Calls, Emails

Statistically Orchestrated Approach to Solve this Problem: The above issue arises due to a disconnected and isolated understanding of the demand management which in turn is a result of system latency, ease of use and recency of data. What we need to do is to implement a statistical process to evaluate the actual inventory based on the controls internal to the system, The forecasts received from the field should not be used directly, but to be taken through the filter of data feature engineering, normalising, trend analytics to define and deliver the inventory control charts for a given period or instance of time segregated as a geographic, demographic or industry based attribution.

Fundamental solution lies in moving from a simple Min-Max based forecasting to a better statistical model like Auto-Regressive Moving Average (ARMA) model (Constantino, Francesco, et al. "Exploring the bullwhip effect and inventory stability in a seasonal supply chain." International Journal of Engineering Business Management 5.Godi?te 2013 (2013): 5-23), which helps to remove the bull whip effect from the forecasting and providing a better and reasonable triggers for production planning and control.

The process flow to follow will be as follows:

By attempting the possible ways and means to reduce the unnecessary cushion and padding in the forecasting trails will help the modern India enterprise to cut down the cost tied up in the days of inventory attaining the cost competitiveness and moving this benefit up the value chain to deliver the same to the customers.

Wish to explore more on how to make logistics a competitive advantage? write to us at [email protected] , At BlackTrunk Logistics Pvt Ltd; our approach is to deliver the customised solution to the logistics and supply chain Achilles heel and helping the enterprise build a Logistics 4.0 roadmap and the long term competitive advantage thereof.