When the customer’s tolerance time is much lesser than the supply lead time (from production to retail shops), it is imperative to have stocks at the point of sales. Not having availability when the customer comes to buy at the store/ shop leads to lost sales. As retailer does not want to lose customer he will actively sell ‘other’ brands, and if the customer’s experience with the ‘other’ brand is satisfactory, may lead to a long term loss of sales for the company. Usually customers want to justify their purchase and can actively market the ‘other’ brand leading to further loss of sales. Hence even infrequent stockouts at the shops can lead to a significant loss of sales.
Many sales people believe that having huge inventory at the point of sales or at the distributors can avoid unavailability. As a matter of fact, in cases where many SKUs are to be managed, high inventory at distributors and retailers leads to unavailability. A contradiction? Combining the facts that sales forecasts are not accurate as we move towards the retail end and that Sales push available stocks (to meet sales targets) the inventory at the distributor and retailer, though high, is mismatched. Consequently the limited cash of these trading partners is stuck mostly in SKUs that are not going to sell immediately while those selling immediately are stocked out. Either the distributor waits for cash to be unlocked or the company does not supply due to credit limits. In both cases the unavailability is ‘created’ though the company has enough stocks in its warehouses.
Additionally introduction of new items or promos can be delayed due to high stock in the pipeline. Having high stocks is not favoured by the finance people as it leads to increase in costs (warehouse space, carrying costs, damages, writeoffs etc) and requirement of more than desirable capital. Most companies would have more than 4 months of stock in the pipeline from plant to the retailers. This stock in the pipeline is the company’s liability as further sales of the company are restricted by this stock in the pipeline.
To summarise to protect sales i.e. avoid unavailability the company needs to have high stocks, while to control costs the company needs lower stocks. But both ‘control costs‘ and ‘protect sales‘ are critical needs to have a profitable growth.
Management frequently swings between the two actions (high inventory and low inventory) depending on the pressure on the critical needs – protect sales or control costs. In many companies this is the way to manage the situation.
Many would consider this situation very complex. In Theory of Constraints, no system is considered complex. If the cause and effect linkages of all entities in the system are understood then the system becomes simple. Once the cause and effect linkages are known, the solution is usually very simple.
Direction of solution:
Obviously keeping high inventories is not a favoured direction due to the damages it generates in terms of inventory carrying costs, capital locked and unavailability. Distribution and Sales want higher inventories to protect sales because they believe that forecasts are not accurate, replenishment times are too long and supply is unreliable. Usually companies try to address this situation by seeking a tool to have more accurate forecasts. Also, managers believe that they cannot reduce the lead times drastically and improve supply reliability without additional investment (additional warehouses and inventory) and a huge effort in the company (as it involves many depts.)
As it is impossible to have accurate forecasts (one cannot control or predict external factors such as competition activity weather, supply unrelaibility etc), we will not try to improve it. Instead we will try to reduce replenishment lead times and reliability of supply while releasing investment (reducing inventory). ‘Theory of Constraints‘ thinking provide the necessary solutions.
Elements of the solution:
If the next stocking point (regional warehouse, Distributors, retailers) is replenished only what is sold by it, the stocking point can have near 100% availability. But if the initial inventory of the SKUs at the stocking points is inadequate (to cater to any demand in the supply lead time), then this replenishment cannot guarantee 100% availability at this stocking point.
As forecasts are never accurate and demands fluctuate, there is no point in undertaking an extensive research to determine the correct inventory levels for each SKUS at each stocking point. If there exists a reliable mechanism to change stocks levels according to the trends in demands or changes in the supply conditions, then a good enough sarting stock is sufficient. This good enough stock levels is determined by-
- Good enough starting stock level at a stocking point= average dispatch (or sales) per day x Total Replenishment Time (TRT) x a factor of safety (to account for variability in demand and supply)
- TRT is the time taken from the dispatch (sales) of the SKU till it is received from the delivery point.
Consider the case of a regional warehouse (could be a regional Carrying and Forwarding Agent). By the above formula the inventory to be carried would be very high – in the range of over 45 days- to have high availability, as the TRT is over 40 days. Due to the monthly planning horizon and that the plants produce in large batches (trying to achieve highest efficiency and utilization), the SKUs are generally produced 2-3 times a month. So a regional warehouse could receive material 2-3 times a month. So from the time the monthly order is placed till the material is received the time elapsed is over 30 days. Adding some factor to account for variability in demand and supply reliability, the safe inventory will be over 45 days.
Using TRT of 45 days does not reduce inventory. As average demand is an external facor, the only parameter to be reduced to have lower inventories is TRT.
The key to reducing TRT is to understand its composition. TRT is made up of-
- Order lead time (OLT), (time till the next order is placed for a SKU since the last order)
- Production lead time (PLT)
- Transportation lead time (TLT) which accounts for actual transportation time and the time for waiting for full truck batches.
Interestingly, if regional warehouses or distributors place orders once a month for an SKU, the OLT is 30 days. The graph below illustrates the inventory level of regional warehouse ordering once a month from the plant. This is also the case of a distributor ordering directly from the plant.
If the regional warehouse orders every day, the OLT will be reduced to 1 day- a huge reduction! With today IT technology capabilities, this data can be easily transferred to the earlier delivery point.
If the supplies to the regional warehouse take place from a plant warehouse which has 100% availability, the production lead time can be eliminated from the TRT.
The plant warehouse receives daily sales data from the regional warehouse and has near 100% availability of all SKUs, then it can fill up trucks with full load with an assortment of SKUs, rather than, as in the existing system with no plant warehouse, wait for the large batch to be produced. This increases frequency of trucks leaving plant warehouse (not number of trucks) thus reducing the TLT by more than half.
To summarise the TRT reduction:
The OLT reduced to 1 day from about 30 days, PLT reduced to zero from the normal 15-30 days, and TLT reduced to less than half. The effect of these reductions on the inventory levels and availability is illustrated below in blue-
The safe inventory to be kept at the stocking points has reduced to about 10-15 days from the 45 days earlier.
But the concern with the above solution is that the addition of the plant warehouse, which did not exist before and is supposed to maintain 100% availability to cater to demand of the whole country, will increase the inventory in the distribution pipeline substantially. Currently the plant supplies directly to the regional warehouses.
Not so! The inventory to be maintained at the central warehouse / plant warehouse = average demand per day on it x production lead time x factor for unreliability of the plant The plant warehouse (or central warehouse) is an aggregation point where statistical fluctuations even out. Not all regional warehouses will have peak demands at the same time. By principle of aggregation, maintaining half the average demand of all the regional warehouses at the central warehouse is sufficient to guarantee availability. TOC solutions reduce the production lead times by half. So, for the plant warehouse we have demand per day reduced by nearly half and production lead time also reduced by half, ensuring that the inventory required at the plant warehouse is not more than 15 days. Incidentally having a plant warehouse decouples the plant from the fluctuations of demand in the market, thus reducing emergency orders on the plant, which further reduces the production lead time and the unreliability. Hence the factor to account for unreliability is also low.
Accounting the already reduced inventory at the regional warehouses (15 days!) and the reduced requirement at the plant warehouse the total inventory in the system upto the regional warehouses, inspite of the additional of central warehouse, has reduced from 45- 60 days to 30 days.
Usually most distributors are just 1-2 days away from the regional warehouses. With the model described above, they can have near 100% availability of the full range with inventory of about 10-15 days.
Their inventory turns will increase to 24 from the earlier 6
. For a distributor with 4% margin this will mean a Return on Capital Invested of over 50%! (Currently they have to be satisfied with about 12-16%) Fine, but then demand is influenced by many external factors such as competition activity or inactivity, seasons, festivals, events etc, and can change significantly. In such cases the original stock levels will be insufficient to cater to the growing demand. The Buffer Management system of TOC manages these changes effectively.
Buffer Management system
: The initial stock level is designated into three colors- red, yellow and greenfor each 33% band. Stock at a level lower than 66% of the designated stock level will be in red.
With increasing demand trend, the stock level will continuously be in red as the supply is insufficient to cater to the rising demand. If the stock remains in red for a period of time equal to the replenishment time to that stocking point, the stock level is changed by 33% ie one color band. Any further change in stock level is carried out only after observing the stock levels after the addition (33%) material has already arrived at the stocking point. Similarly for decreasing demand trend, the stock level is reduced by 33% if the stock is always in green. For steep changes in demand such as an event (Diwali) or a season (air conditioners) the levels will have to be changed manually, and later managed by Buffer Management during the season. As the supply is only according to what is sold in the next stocking point (distributor or retailer), and buffer management manages the upward and downward demand trends effectively, there is obviously no need for a short term forecast!!! Long term forecasts are required to manage only the long lead time raw materials such as imports.
The model described above provides near 100% availability at all stocking points while reducing the inventory in the system to nearly half. The improved availability increases the sales, while the distributors and retailers experience significant increase in ROI. A WIN-WIN for all partners in the chain.
So the magic pill: Stop forecasting, hold back inventory rather than pushing down the chain and supply according to consumption.
This comprehensive solution, with all its elements has been implemented in quite a few distribution companies (consumer goods companies, autoparts company, fashion goods etc) in India by Vector Consulting Group. The results- more than 30% sales growth at half the inventory levels, in a year.