To teach lot size optimization I like to use an example that is close to home. Suppose you live in Georgia and there is only one ATM machine in the state. The single ATM is located in a small, remote town in south Georgia (Are there other kinds?). The ATM is only open during the last week of July. For whatever cash you need you endure a trip down heavily congested country roads through swarms of gnats in humidity so thick you will need an umbrella in heat so intense you will think you are in a sauna to stand in lines so long you will think you are at Disney World. One last thing. When you finally get your turn at the ATM the fee to withdraw your money is $1,000. How much money will you withdraw? Yep. All of it. While it is not in the bank, it is not earning interest. It is likely to get lost or stolen, and if you are like me you are much more likely to spend it.
The sum total of the hassle, pain, and literal withdrawal fee equate to the transaction cost. In general, the higher the transaction cost, the fewer times we want to endure the transaction.
In manufacturing and production contexts, the transaction cost related to lot sizing is the cost, hassle and time of setting up or changing over a production line. The higher the cost, the longer the time, the greater the hassle, the fewer times we want to execute the transaction. So, when we get the line setup, we are should wisely run it for a while. The result may be a lot of inventory. (No wonder it’s called a lot size.)
In the purchasing and procurement context, the cost related to the time, telecommunications, planning, and execution of a purchase order is the transaction cost. As was the case with production, the higher the cost, the longer the time, and the greater the hassle, the fewer times we want to execute the transaction. So, when it comes time to place a purchase order, we are going to order a lot.
Now, let’s run the south Georgia ATM tape back and fast forward to another time. Suppose there is an ATM machine within arm’s reach wherever you are. It is open 7x24x365. There is no “withdrawal” fee. (Amazingly, it’s free to access your own money.) Now, how much will you withdraw each time? Yes, just enough for the next few minutes or until some cash is needed.
In manufacturing, if it’s free and requires no time to setup or changeover a line, we can afford as many setups as we wish. Manufacturing lot size inventory in those cases should be minimal. In procurement, if it’s cost and hassle free to place a purchase order, we can afford to place as many as we like. Procurement lot size inventory in those cases should be minimal.
Transaction costs should go a long way toward determining the size of the transaction. Unfortunately, not many supply chain organizations can tell you the true cost of their most important supply chain transactions; setup costs, changeover costs, purchase order cost, freight bill payment costs, transportation setup costs, etc.. As a result, lot sizing is often overlooked as an opportunity to improve inventory and supply chain performance. Lot sizing has become a lost science in supply chain optimization. Even the EOQ has been one of the babies thrown out with the bath water.
In an attempt to re-institute lot sizing in supply chain strategy, we developed and execute lot size deviation analyses as a part of most supply chain assessments. We typically find that lot sizes are off in one direction or another by 100% to 300%. An example from a large food and beverage company is in Figure 3.18. Note that the lot sizes for 86.5% of the SKUs were undersized; to the tune of 50% of the optimal lot size. Once corrected, total supply chain costs were reduced by more than $10 million; the majority of those savings accruing from higher manufacturing productivity .