With all the recent discussion regarding production issues and Cirrus and the recent layoffs, I though it would be fun to propose a solution that could be implemented quickly and possibly speed production. I am not an altruist so my reasons for this post are only to get my plane as early as possible.
First a caveat… I have yet to see the production facilities in duluth. However, I own several manufacturing companies with labor intensive products and know a bit about manufacturing and what I am going to discuss.
In the 1980’s a gentleman named Dr. Eliyahu M. Goldratt ( a physicist by training ) developed a concept called the Theory of Constraints. This theory, when properly implemented, has allowed many manufacturing companies to dramatically improve manufacturing cycle time with a commensurate increase in profitability. In some cases, the improvements were dramatic and very fast and required no additional capital.
In one of my facilities, we are implementing this theory to increase throughput 86% while cutting assembly space in half!
How does the theory work?
This post will be long enough without going into great detail about the theory but here is the simplified version.
The first part of the theory is that we have to look at Cirrus Design, as a whole, is a system. It is a system designed to manufacture money. In this case the goal is done producing airplanes, but the whole purpose of everything Cirrus does is to make dollars. Therefore, Cirrus Design is a system to make dollars.
Cirrus makes dollars only when a customer stands on the factory floor in funny looking sandles and is handed a set of keys. All the steps Cirrus must take to get to this point are the processes within the system.
The Thoery of Constaints (TOC) suggests that all systems are similar to chains. Each chain is composed of a variey of links (different processes) differing in strength and capability. For the purposes of our discussion, we will concentrate on the actual steps of producing an airplane and ignore the administrative details. A chain with differing strengths in its length will only be as strong as its weakest link. The strength of the whole chain or system is not an average of the strengths of each link.
If you think of all the steps that are needed to produce the SR20 / 22, you can imagine that some are easier than others. Each step represents a link in our chain and if we think of all the processes, we can conclude that some processes are capable of producing more parts in a period of time than others. However, the total system will not produce airplanes based on an average of its processes. It will produce airplanes at exactly the same rate of its weakest process, whatever that may be.
So if we have five steps in production, we could say that step 1 can produce 39 parts per hour, step 2 can produce 23 parts per hour, step 3 can produce 15 parts per hour, step 4 can produce 27 parts per hour and step 5 can produce 32 parts per hour.
Now the question is, what is the overall capacity per hour of our five processes. That’s right, 15 units per hour. Regardless of how good the rest of our processes are, all we are going to get is 15 units per hour out of this system. This is kind of common sense, right?
The problem is that most manufactures try to optomize the whole system to make more parts. The reality is that you only need to identify the constraint in the system and strengthen it. Once you have done this you optomize the whole system to the next constraint. In our example, this is step 2 at 23 parts per hour. And this can be done quickly and simply by focusing on the constraint.
The key is in identifying the constraint. In most manufacturing plants the constraint has a bunch of work piled up in front of it but there are more “scientific” ways of identification.
The idea is to keep breaking the constraints until the constraint is external (market demand).
Now this sounds simple but here is why it is difficult for most of us manufacturers to adopt.
We live and die by costs. Labor is a big cost in my facilities and I imagine it is expensive at Cirrus. So what do we do… we make sure that all of our workers are staying busy all the time. Well in our example above, the whole system is contrained to 15 units. Therefore having the other processes optimized for “efficiency” (an employee or machine making the most parts per hour) is dumb and leads to high inventory costs and storage requirements. If we are not disciplined, we will spend a lot of time and money strengthening links in the chain that are already okay.
If the constraint cannot be fixed (cure time for the wing takes 8 hours (or whatever) then we should be setting our production plan to the speed of the constraint ( build one plane every 8 hours)and then sub-optimizing all the other processes before and after the constraint.
Therefore step one is to identify the constraint and try to strengthen it (focus our effort). We keep doing this until we reach an internal constraint we cannot stregthen, or find our constraint is external (market demand). If our constraint is internal and cannot be strengthened, then we set the entire plant to the pace of this constraint, knowing that many processes before and after the constraint will be sub-optimized. This is why cross training people is important because you can move them where needed. Just make sure the constraint is always working and is never on break.
A quick note then I’ll close this post.
Most constraints are policy constraints and not a physical process constraint (wing curing time for example). A policy constraint is often the result of standard GAAP accounting rules for manufacturers that lead to cost analysis per process. As mentioned before, just because a process can produce 39 units per hour does not mean it should if the system constraint is less. To do so is wasting resources that cannot be used.
Rather, the accounting for the plant should be based on throughput. In other words, how many dollars the plant makes every time a part makes it through the constraint. These are the only dollars that matter. We throw away the whole notion of inventory value added for parts made in processes that don’t yield dollars. Remember, dollars are only made when the customer is in sandles with keys being handed over. This only happens when a part makes it through the constraint. Focus on the constraint and the dollars will be there! (As you can imagine, if the whole throughput rests on getting the part through the constraint, then having the constraint down is way more expensive than GAAP accounting would figure it as being)
What should Cirrus do. Get someone to their facility who knows TOC and get the constraint identified. Once the constraint has been identified, work like hell to optimize it while making the rest of the plant dance to the beat of this drum (the contstraint schedule). Perhaps the folks at Cirrus should pick up a copy of William Dettmer’s “Breaking the Constraints of World-Class Performance” (Amazon.com) I have a well worn copy I would be willing to lend.