Optimising operations over the supply chain can be a daunting task. Puneet Kulraj shares a few simple solutions to the problem of unpredictable stock-outs Managing the supply chain of a distribution company, be it a consumer goods company, a retail chain or a spare parts distribution centre, is extremely challenging. Managers face frequent conflicts in handling the day-to-day decisions for the supply chain. The challenge Take, for example, the chronic conflict between the Sales and Logistics teams about inventory. While the Sales team always wants higher inventory to protect sales, Logistics and Finance want to limit inventory to control costs. Sales is more than willing to start a discount scheme and push out inventories to counter competition, while Finance is wary of such drops in product margins and views this as a great loss. Production wants schedules (based on forecasts) to remain stable, while Marketing wants production to be more flexible to the changes in market requirements. Sales wants more budgets allocated for marketing and advertising expenses, while others may insist on the growth of sales to fund the extra allocation for marketing and advertising expenses.
 These conflicts manifest in seemingly contradictory supply chain issues such as significant stock-outs despite high overall inventory (inventory turns of around 3 or 4), price pressure from supply chain intermediaries while the price for the end consumer remains unaffected or new products introduced when the old ones still clog the pipeline. The problem is further aggravated for high-tech products, where the total inventory in the pipeline for most companies is usually much greater than the lifecycle of the product itself. This leads to significant price discounts and a subsequent negative impact on the profitability of these companies. The same problem is seen with fashion products. A large number of SKUs need to be available a long time before the fashion season begins. There are stock-outs on about 20ñ30 per cent of the items (those that sell well) during the initial weeks of the long fashion season. Towards the close of the season, there are many slow movers which have to move to the 'factory outlet' for discount sales. The ramifications of dealing with such supply chains are significant for the retail shop. A retail shop is always constrained by cash and/or space. Most of the shop inventory is skewed towards the slow movers. Since the slow movers block cash and space, more sales efforts and space are allocated towards the slow movers. This denies fast movers the opportunity to clock more profitable sales. At the same time, shortages of the fastmoving SKUs do occur. With cash tied up in inventory, the ROI for a retail shop is often less than desired. With such chronic conflicts, significant growth in sales (about 30 per cent over the previous year) is never targeted in the annual business plan because many believe such rapid growth in sales will invariably come at the cost of very high growth in expenses, lower margins or high inventory. The targets are grudgingly set at less than 10 per cent growth over the last year in many distribution organisations. As an answer to this, a manager of a large distribution company once remarked, 'We can set and meet any ambitious target, if we are able to be 100 per cent accurate on our sales forecasts. An accurate forecast will ensure we make all the right SKUs, and distribute it to the right location of demand and do not feel the pressure to drop prices since our inventory matches the forecast.' Through the looking glass An accurate forecast looks like a good direction for a solution. Many distribution companies are struggling hard to improve the accuracy of their forecasts. Many have invested in expensive software tools to improve the forecasts. Despite all the investments, the problem remainsóforecasts are not nearly as accurate as one would want them to be.
However, the fundamental question remains: Can forecasts ever be accurate? We may be able to provide a 'reasonably' good forecast for a product at the national level, but forecasting at SKU and location level for a long horizon is as reliable as weather forecasting. Chaos theory validates the fact that it is almost impossible to accurately predict the outcome of a chaotic system, such as the demand in a market. A small change in any demand variable can lead to disproportionate changes in the outcome, thus making it difficult to predict the outcome. For example, when a competitorís supply is affected in one location, leading to a sudden surge in demand, scarcity- driven purchases boost sales, or there is a sudden surge in demand for a fashion SKU after a local celebrity is seen wearing the product. The sales at the SKU and location level are highly fluctuating and unpredictable and therefore it is impossible to predict the impact of all variables on demand. Any attempt to bring sanity at this level is a futile exercise.  The Solution: Simple Yet Powerful The starting point of building a good solution is to go back to the basics of inventory management. One needs to understand that inventory is kept for products where the customer tolerance time is much less than the time to produce, hence there is a need to keep inventory so as not to lose the customer.
The inventory at any location has to account for the following variables: Replenishment lead time Demand during lead time Variation in supply time Variation in demand Of all the above variables, replenishment lead time is the most important as it impacts all the others. So if one wants to improve the forecasting, it is important to focus on one variableóthe replenishment lead time. If we are able to reduce the replenishment lead time significantly, we can manage with much less inventory. Forecasting accuracy will improve significantly with lower lead times. To reduce supply lead time, one needs to understand the components of the replenishment lead time. The components are: Order lead time (time till an order is placed for a SKU) Production lead time Reducing the Order Lead Time The order lead time is a significant part of the total lead time. The min/max ordering system increases the ordering lead time, as one has to wait till the level reaches the reorder point before placing orders. Many organisations take orders from distributors once or twice a month per SKU, even though they might be required to take many orders throughout the month. Can the order lead time of an SKU be reduced? Yes, in the era of interlinked computers and Electronic Data Interchange (EDI) systems, a daily ordering plan per SKU can be adopted. This does not mean that the supplier has to ship out more frequently with partial loads.
If there are frequent shipments, each shipment will now have a larger assortment of SKUs rather than a single SKU. Managing with Lower Inventory Now that the supply lead time has been reduced significantly, one needs to stock enough to cover any demand that may occur during the supply replenishment time. Then next step is to focus on fast replenishment to the actual consumption from the stocks. The need of forecasting just goes away. The entire supply chain just reacts to a very objective dataconsumption.
For every item, at every location, a target inventory is set based on ëparanoidí consumption. During replenishment time (the maximum forecasted demand during the replenishment time is factored by the fluctuations in the replenishment time). The stock is replenished at the pace of the sales. Managing Exceptions with Buffer Management Even though supply is based on consumption, there is a chance that the inventory might fall down to dangerous levels between two supply runs, which can lead to stock-outs. The target inventory is a planning decision. We need a system to manage such exceptions during execution.
Buffer management is an execution control method that provides priorities based on the actual consumption of the buffers. The target level of every item, at any location, is a buffer. Buffer status measures how much of stock does not reside at the location in comparision to the target level. It is defined as the percentage of the target level on hand to the set target level. What is missing from the on hand stock should be somewhere in the pipeline from the source to the target. When the stock at the target is more than two-thirds of the target level, the buffer is considered to be 'Green'. This means too much stock. When the stock at the target is between one-third and two-thirds of the target level, the buffer is considered to be ëYellowí, which means that the stock level is adequate. When the stock at the target is less than one-third of the target level, the buffer is considered to be 'Red' which means there is a real risk of running out of stock. Changing the Target Levels If the stock stays continuously in the Red throughout the replenishment period, changing the target levels might be necessary. Similarly when the stock stays in the Green for multiple replenishment periods, it is time to look at reducing the norms. Dynamic buffer management helps in aligning stocks based on demand situation per SKU per location. The decision is much better than current practice of just arbitrarily cutting stock norms. When one arbitrarily cuts the stock norms, people usually reduce the number of fast-moving items, thus leading to more stock-outs. Dynamic buffer management prevents such ad-hoc decisions.
 Benefits of Aggregation—The Plant Warehouse In most push-based supply chains, the bulk of the inventory is close to the demand point' where demand fluctuations are most erratic. A sales target based on primary sales pushes inventory close to the demand point. This leads to stock-outs of an SKU in one location which may be available in excess in another location. Many organisations have tried to solve the problem by inter-warehouse transfers, but the rising costs of logistics and the temptation to hold on to inventory at local points has made this solution ineffective.
We can solve the problem by having inventory at a point close to supply, where demand is flat and stable. A plant warehouse with the bulk of the inventory feeding the regional warehouse should solve the problem to a large extent. The plant warehouse will decouple the regional warehouse from production fluctuations. The inventory at the plant warehouse will account for production lead time, while the regional warehouse needs to stock only enough to account for the transportation lead time. The plant warehouse will reduce the overall inventory levels in the supply chain, as fluctuations at the central warehouse level are lower than those at the regional or the retail level. The demand points are free from production fluctuations, so they can manage with a much lower inventory than they currently have. The supply to each inventory location will be based on consumption. For example, the plant warehouse will supply to the regional warehouse based on consumption; similarly, the plant will produce based on consumption in the plant warehouses. The Expected Benefits Stock-outs should drastically reduce as replenishment reacts to pace of sales much faster than before. In most environments, a sales jump of around 30-40 per cent is visible after implementation of the replenishment solution. The inventory drops drastically while the ROI of dealers goes up significantly as they manage their business with a much wider spread of SKUs without any stock-outs.
The author is a founding Director of Vector Consulting Group. |
