SMED is a simple concept really. It sounds a lot more complex than it actually is, really. The reality is it’s just looking at a process and determining a better way of doing it based on the current scenario – focusing on a few points, moving some things around for an optimal time while focusing on changeovers. That being said, we will really dive into what it is, in a second. The plan is to just show by example – make it applicable – and hope that you can find some use for the concept in your organization. If you work for a larger organization, and if it is a manufacturing company, you are most likely performing some sort of this but may call it something different.
So, prior to constructing any article, I do a little research. Not only does it take what I already know into consideration, it allows me to get thoughts from others as well as some of the history. Many times during my training(s) the history was left out. However, I did find this tidbit interesting. So one of the initial concepts for SMED or how to optimize changeovers, was to change the lot size being ran. A changeover, for those not in manufacturing, is an event that can range in complexity or time where machinery, staff, materials need to be exchanged due to the production is being changed. So we may be making cars and then going to trucks. Or perhaps (depending on what type of manufacturing) we may be going from running a blueberry pie to a cherry cobbler. Different types of products require different materials, and many times completely different machines or staff to operate. Back to my original point, the lot size… One of the earlier concepts was to adjust the frequency of changeovers by changing to run time of the current product or batch size. So instead of running 5 hours and taking a 1 hour changeover, then 5 hours and another changeover we may run 10 hours on a single product and take 1 changeover. 1/5 vs. 1/10 changes the amount of time available you are dedicating to a changeover by rearranging schedules for larger runs. However, that is not the trend of modern manufacturing – where we are in higher demand to be agile and run lower volumes with more changeovers as products are in smaller demand through niche markets or JIT (just in time) business needs. JIT, for those not aware, is the concept of making the product just as it is needed for the next part of the process – not sooner – or later which controls WIP (work in process). WIP = inventory. Inventory is tied up cash flow and risk if something were to happen while waiting to be processed. Sorry, a lot of manufacturing lingo…
So, back to SMED. Let’s say we are moving to dessert from the main course at a fancy restaurant. The customers need to be kept happy – and the longer it takes to get them in and out the table is tied up so no new guests can come in. Not to mention you are paying the employees during this time whether they are selling more food or waiting on the same guests to finish the current meal. This way it is a simple example, and maybe you can take the simpler example and extrapolate to a process in your office, warehouse, or whatever business you are a part of. There are two types of main processes in a SMED. There are external activities, or tasks which can be completed outside of the process at any time and internal activities, or steps, which have to happen in a certain order at a certain time, or cannot be prepared ahead. We cannot bring the dessert out until the dessert is ready (internal). It has to take place in a certain order. Yet, we can prepare the next plates, utensils etc. at anytime to bring them out (external). Ideally, during this process, we would try to convert as many of the external tasks to internal as possible. The more steps you can prepare or do at any time, the better, and quicker the process can become.
In the example below, instead of each step being performed prior to the next, we combined times. So while the server was waiting for the guests to finish the table was continuously being checked and cleared as able. In addition all the dishes etc. were being prepared while waiting for the guests to finish. Since we knew how many guests – we could assemble that number or settings. If one was not used, we could set it aside for the next order. I am sure we could even create an inventory or prepared settings and the servers would never have to prepare, but pull from the WIP (work in progress). While he or she takes the order the remaining plates were picked up. All of the dessert dishes were then carried to the table with the dessert being brought out. Our bottleneck, or part of the process we cannot really change is how long the chef took to make the dessert. In our little example, we saved 25 minutes. Perhaps that is equivalent of how long it takes for someone to eat a meal, meaning we doubled our turns of that table potentially. Get the picture?