Bathtubs & Co is a manufacturing company that produces bathtubs and jacuzzis to order, mostly to constructors. They decided to implement S-DBR and BM six months ago. Not surprisingly, after choking the release and establishing priorities their DDP (Due Date Performance) crept up from a lousy 55% to over 97%, leading to a 35% increase of sales and more than a 100% increase of profit. Gabriel Díaz, the company’s owner couldn’t be happier and can’t speak enough about how great TOC is. However, his newfound love didn’t last for long.
As sales began to increase so did the percentage of red orders. For several weeks red orders ranged from 20% to 25% and shortly after, just as expected, the plant started to miss some due dates despite the efforts of exploiting their CCR and working 24/7.
Amanda López, the plant manager, decided to increase buffers as suggested by conventional TOC knowledge. This way, she thought, the plant would be able to recover and stabilize. However, the opposite happened. Lead time and work in process (WIP) increased, with a consequent increase in red and black orders. Amanda concluded that the first adjustment had not been enough so she augmented the buffers once again. But, just like before, the situation got even worse. It became so critical that clients started to cancel orders swearing to never buy again from Bathtubs & Co. Gabriel was furious and Theory of Constraints took all the blame. The company stopped the S-DBR implementation and any initiative that even smelled like TOC. What do you think happened? What did Bathtubs & Co do wrong?
Let’s look at the facts. Bathtubs & Co choked the release and established priorities improving DDP and increasing sales; no mystery here. However, when sales increase then more and more WIP will be released to the plant, sometimes much more than what the CCR is able to handle during a given time period. For example, if many clients place an order the same day and the company promises a standard due date, then all the orders will receive the same due date and the same release date. This will cause queue time to rise and orders will inevitably be late. When Amanda increased the buffer what did she do? She just added fuel to the fire. She thought that by releasing earlier she would have more time to deliver the order but if the CCR is already heavily loaded as it is, releasing with more anticipation won’t help. As a matter of fact it’s worse because more orders will be competing for the already scarce CCR and non CCR capacity.
What should Amanda have done differently? First, she should have followed the S&T’s sequence rigorously. Instead Amanda decided to adjust buffers before implementing planned load which was a huge mistake. To understand why we must understand the difference between traditional DBR (sometimes referred to as Full DBR) and S-DBR. In traditional DBR, the drum is the CCR schedule and there are two buffers. One is the CCR buffer which protects the CCR schedule. The other is the Shipping Buffer, designed to protect the due dates. By sheduling the CCR, calculating safe due dates based on the drum, and releasing WIP accordingly, we ensure that we’re not releasing more than what the CCR can produce. But in S-DBR we only have one shipping buffer and no CCR schedule. The drum is just the list of orders with their respective due date. This means that in S-DBR we are releasing RM solely based on the orders’ due date, irrespective of the CCRs capacity or load. In other words, we are subordinating entirely to the market. If demand is less than the CCR’s capacity, there will be no problem. But when sales trigger more and more orders to be released, we’ll be in trouble. This is where planned load comes into play because in the absence of a CCR schedule we need a way to quote due dates based on the CCR’s load. This will smoothen the load on the CCR and limit how much WIP we allow in the plant.
S-DBR and planned load are not separate solutions; even though they are two separate tactics, they make part of one unique solution and must go together. This doesn’t mean that you have to implement planned load from the start, but it does mean that only after implementing planned load should you move to ajusting buffers (entity 4.13.1. under entity 3.1.3. – “Accomodating Growth” in the standard RRR Tree). The sequence embedded in the S&Ts is not arbitrary. Not following the sequence can cause serious harm to a project as the above case illustrates.
