Inventory modelling strategies for Timken India

Abstract

Timken India is one of the largest bearing manufacturers in India and has a complex supply chain with globally distributed manufacturing facilities and distribution centres. It mainly serves the automotive OEM and process industries, with a smaller percentage of sales in the after-sales market. It maintains a responsive supply chain which aims at maintaining buffer inventory to deal with demand or supply uncertainty and is located close to the automobile manufacturing hubs in western, northern and southern India. The company faces dynamic demand for products that results from changes in market conditions and an ensuing challenge is of maintaining inventory control which meets the required demand at a minimum cost. The primary manifestation of the inventory challenges for Timken is evident from the low inventory turns in comparison to industry competitors. Though Timken has a very comprehensive inventory model, interactions with company’s supply planners reveal that high inventory levels are being maintained to cater to the volatile demand from the market. A study of the inventory model suggests that classification done through the model does not take into account the profitability of the particular SKU – it depends only on its demand profile. The problems with demand management were validated by studying the inventory holding patterns and the demand fulfilment resulting from it. A small sample of products was individually considered to understand the ordering patterns with respect to the present and past sales, and the current levels of inventory. The next part of the study focussed on analysing and identifying the changes which could be incorporated in the inventory policy. Given that there is excess inventory and variation in demand fulfilment, the holding costs were analysed to find the trade-off between continuing with high service levels and the cost incurred in maintaining inventory to facilitate such service levels. To understand the relation of profitability with demand volatility and overall volume, demand segmentation was done to group products in four quadrants. With certain assumptions the results show that a small number of products contribute to disproportionate amount of profits. To ensure inventory rationalization, the planners can utilize Pareto analysis to classify the categories of inventory. As such simplistic classification may not satisfy the various criteria utilized to categorize inventory, a multi-criteria classification technique has been discussed. A large number of SKUs in the sample data display sporadic demand and hence Poisson demand distribution was considered to be more suitable than normal distribution in such cases. It has been shown that considering Poisson distribution can reduce the inventory holding for such products. Based on the data shared by the company and the resulting analysis, recommendations for improved inventory management have been made. Demand segmentation of the product portfolio based on required criteria should be done to indicate the service levels required for different classes of products. The service level can be optimized and the safety stocks lowered through better forecasting. The inventory control techniques should also be modified as the product range indicates large number of products with sporadic demand. Finally interdepartmental coordination between functions should be improved to increase visibility in supply chain and reduce overstocking due to ineffective communication. An analysis of cost reduction through the implementation of the recommendations could not be conducted due to the lack of cost and profitability data.