Tom Inglesby: John, you are credited with the concept of Demand Flow Manufacturing. How did you come up with the idea?

John Costanza: It goes back about 15 years, to when I was with Hewlett-Packard. I had an opportunity to take a look at manufacturing in general, from an engineering standpoint, and to see if it was possible to design a new form of manufacturing. We were looking for a method that would be mathematically based — just like we would design a product — and based upon response and speed as opposed to scheduling and tracking. The thought at HP was, if that could be done, it would be pretty incredible. Better, if you could do it and formalize it, you'd have a huge leg up in the world of systems. I set about to do exactly that, to take a look at design and manufacturing response. The result was Demand Flow. The key word is demand. How you drive a factory to demand is very interesting, especially with demand patterns that can change minute by minute.

That's where the whole thing came from. The bad news from the Hewlett-Packard standpoint was what I was designing and developing had no resemblance at all to the computer systems that we had at HP. We were based upon MRP methodologies and scheduling, and the methodology that I was working on pretty well eliminated transactions. If you think about that, if you eliminate transactions, the true value of the computer starts going away rather fast.

So you're really not looking at it as a computer system, but rather as a philosophical concept.

Correct. It was like the guy who worked for Exxon that developed the 600 mile per gallon carburetor. It was good news to the world, but not good news for Exxon. I had a great concept, but it wasn't necessarily good for the computer world.

There was no software in existence at that time that could be modified to handle Demand Flow?

None at all. If you look at them, they are two totally different methodologies — scheduling vs. a demand-driven flow process. In the scheduling world, we design bills of materials, and we use the bill of material to actually drive the process. By that I'm saying if you design subassemblies, then that's what manufacturing builds, subassemblies. The whole process is controlled by the bill of material.

In a flow environment, we use mathematical calculations called operational cycle time to define targeted work. That's what defines work content — your operation cycle time calculation, not your subassembly. We design it one time and that's established to be capacity. Then you leave it alone. In a flow environment, driven by demand, operational cycle time defines work; in a schedule methodology, subassemblies and fabricated parts define work. The bill of material in a flow environment is just a pile of purchased parts. The bill of material in an MRP world is multi-level hierarchical design for scheduling. The two foundations are totally different.

I understand how Demand Flow could work in a job shop, make-to-order environment. How does it work in a company that's making to stock or making products that go out in mass quantities?

The problem is that there's a belief that a company must be either a scheduler or a flow manufacturer, driving to demand. What if you only knew flow manufacturing? Does that mean you can't build products that are make-to-stock? Does that mean you can't build products make-to-order? Of course not.

Would that require having two environments?

No, not at all. A flow manufacturer could care less about scheduling. A flow manufacturer is going to be driving in the direction of demand just as its volume mix changes every day. You design your flow line for a mix of products, not as a dedicated line.

So if you are in the nail and screw business, for instance, you can make 50 million number 10 whatever, but you can respond to a request to have a copper-plated number 16 if it's needed.

Absolutely, that's what you design your lines for, to go from 8's to 6's in the direction of the demand. And you design a mix-model production line up front, knowing that you're going to have to change volume mix based upon demand every day.

Is there any type of manufacturing that this doesn't work with?

No. It's a different methodology of building products. It will work with any product anywhere in the world at any volume. Doesn't make any difference. This is why the systems people are now being forced to listen to us. They ignored the concept for a long time; and now they can't ignore it, the noise is too big.

It works just as well in a process industry?

It actually works better in a process industry because it's a natural process to begin with.

Give me an example of how it can be used in a process environment.

Let's look at the companies that make Band-Aids, fingernail polish, insulation. There's typically an upstream process of mixing, or recipe building. When we do DFT (Demand Flow Technology), we stop focusing on using large vats for our mixing. When you have large vats, all you can do is one thing; and you have to process the entire vat before you mix another recipe. Instead, we want to have smaller vats and multiple products in process at a given time. Then you can go from product to product in a very short time.

Rather than buying big monster vats like the industry has been all these years, we go to some smaller pieces of equipment, multiple pieces that are a lot cheaper and easier to maintain. That allows us to go mixed modeling. We might have four different shampoos in the hopper at one time as opposed to one. We change the way we design. And remember, at the end of the process, you always have to package the product in some way, shape or form. You also have to design the packaging for mixed modeling. If you're making insulation, you can go from 24-inch pads to 16-inch pads back to 20-inch pads. We design the packaging to be very flexible to go from product to product. That's part of mixed-model line design.

It's a change in the way you look at manufacturing, not just the way you push product in or out of the production line.

You definitely have to look at it differently. Look at it from a market standpoint, if we're going to be responsive to demand today, we're going to change volume mix every day. The days when we can just schedule a monthly quantity of one product, get that out, then go to the next monthly quantity, that's come and gone. Customers are not so loyal anymore. The competition is a lot more fierce. If you don't have your product there, the customer will buy somebody else's. I as a customer am not loyal to any brand name, and neither are you.

How do you determine demand? Is it demand coming down all the way from the end user or demand from your customer upstream one notch?

You can only go one level up. You're at the mercy of the guy above you. Remember this, he creates a demand pattern whether you like it or not. He is a customer. Whether he's right, wrong, brilliant, not brilliant, he is what he is. You can only take the data that's given to you. If you think they're totally off base, you might want to go talk to them; but they are the customers.

There's a growing trend saying customers are not always right, especially if it's going to cost more to satisfy that customer than to make the product.

You can look at it that way. They'll always be the customer, but you may not want to do business with that customer.

That's one of the things that Demand Flow can actually help determine, isn't it?

Absolutely. If this customer is going to require me to carry massive inventories; they won't communicate demand to me; and they're not sharing information with me, but they expect me to pull off miracles, I might just want to tell them to go somewhere else. In a lot of the tier one and tier two automobile suppliers, we're seeing people saying "your volume is so massive, that if I do business with you, I can't do business with anybody else. And if I do business with you exclusively, and you or your business go away, I go away." They're not willing to do that.

How do you balance the idea of demand flow and the old concept of EOQ (economic order quantity) so you can have the benefits of scale?

Don't worry, we're not going to do anything to increase costs. We still want to drive in the direction of demand, but some days there are zero orders and the next day there are 100 orders. I can't lay off everybody today and come in tomorrow and start with twice the work force. The key scheduling of a demand-driven process is to go in the direction of demand. You start changing the volume of mix every day in the direction of demand.

Let's say we're making bicycles in 200 different colors. When we decide to paint black bicycles, we have to clean out a gray paint booth and spray guns. So if I'm painting a black bicycle, I'm not going to paint one, I'll make five. One will get sold, four go to inventory. The cost to make one at a time is ridiculous.

It's not a discussion of one at a time or one less at a time, it's a discussion of flow manufacturing. Ultimate response and economics are definitely a key factor here. In Demand Flow, I try to eliminate all opinion, and stick to a mathematical model because mathematics don't lie. You can use it, and I can use it, and we will get the same answer if we do it the same way.

There's something we refer to as replenishment interval. The replenishment interval, going through a set of machines, can be calculated based upon setup. That would tell you whether you can make those bicycles one at a time or three or four at a time. If you want to get it to one, then you have to get your replenishment interval down. In some places you can't. Economics is still an issue.

Demand Flow sounds like a philosophy more than a theory.

I have no real good feelings about the "philosophers." Eliminate waste, they say. What do you think we've been doing for 20 years, creating waste? Continuous improvement? Well, we've been continually improving up to here, now we get our teeth kicked in. You need to have revolutions. It's a series of little revolutions as you evolve a company from scheduling to flow. It's not shut it all down, revolutionize it. It's a series of little revolutions as you evolve the company.

There is a terminology discrepancy. You use the term Demand Flow. Many of the companies getting into the marketplace are talking about flow manufacturing. What are the differences?

Flow manufacturing is whatever they want it to be or say it is. Demand Flow is based upon a set of principles, a set of mathematics, a set of tools. I'm probably the only person who has formalized the flow manufacturing industry by giving people specific tools, specific techniques. If you talk to someone on Demand Flow, and say, "Well, how do you define a process?" they'll talk about something called a product sink, they'll talk about operational cycle time and balance, all of which are mathematical formulas and foundation.

That's why I moved away from the philosophers. Philosophers come and go. Give me tools and techniques. The thing that I'd say John Costanza did is to give the industry tools. Specific techniques to design a process, specific techniques and tools to balance the process, specific algorithms to take a demand pattern and put it into something that's smooth — that's actually patented by the way — all of which I teach. I believe if you give people good tools, they'll do great things.

We've spoken about the theory of constraints vs. Demand Flow and how they can and can't work together. Give me your opinion of TOC vs. Demand Flow.

I look at the theory of constraints as a very philosophical approach to try to solve a problem. You can't argue with the philosophy, it just has no validity, no foundation underneath it. If you're implementing technology, then what is the first thing you need to do, what is the second thing? I can tell you exactly what they are. You know what the results are going to be. It's not "try it, see if you like it."

I believe you should give people good tools. My objective was to give them those tools. Although I have never met Eli Goldratt, I've listened to him speak. It's not an issue with emotion or philosophy; it's an issue about tools. I just don't think he has the technology to be competitive.

What would you say is the most important aspect of getting a company to look at and adopt Demand Flow?

It's a management issue: commitment. Are you willing to make a change? I think people want the benefit, no question about that. The question is whether they are committed to make the change happen. I see too much management that is in habit of jumping up and down shouting and screaming, but not giving people the right tools and a clear path. If management can't provide leadership, what value are they? The thing Demand Flow does is it gives you a path. If they've got leadership and that path, they will be successful.

But it has to be done a step at a time.

Correct. A step at a time with a path and with leadership. It's easy to go out there and beat the heck out of the direct labor. That's what management is notorious for doing, complaining about the work force. Let's make a change. Let's have tools and commitment. Let's make it happen.

There seems to be a wealth of new companies starting up, especially in the high-technology areas. Is Demand Flow easier to implement when you have a clean sheet of paper?

If you're starting a new company, you have nothing to disassemble. It's much easier and much faster. If you have a company that's currently in place, then you have to change things to go forward. That's more work because you have to back up and change it before you can go forward. The good news is that if you have a company that's up and running, you have experience, you have people, you have knowledge. If you have a new company, you don't have any experience or product knowledge. So, that's actually a little tougher to do.

With an existing company at least you have something to benchmark, something that shows when you are getting some benefits.

That assumes you have choices. My point is, in a competitive world where response is critical, you don't have a choice. You can't use MRP and schedules even if you wanted to because if your competition goes out and does Demand Flow, you're going to be in big trouble. You really don't have a choice.

What are the strategic benefits of Demand Flow, as opposed to the daily benefits?

You get the benefit from the marketing standpoint, which is probably the number one reason why you should do Demand Flow: the ability to satisfy your customers' demands. You get very quick response and you get a good quality product to the customer. The reason you should do Demand Flow is from a marketing and financial standpoint. The manufacturing issues, the productivity and inventory reduction, is a secondary benefit. Even though it's huge, it's still secondary to making the customer happy and to penetrate deeper into existing markets or to open new markets. Do it for 1) marketing, 2) finance, 3) manufacturing.

Since your concept is based on mathematics, there's nothing preventing it from being implemented anywhere in the world.

That's right. Mathematics is a universal language. Just imagine how difficult it would be if you and I were running a store and you had your form of algebra and I had my form of algebra? We'd argue about what the costs ought to be, and the prices on the menu, and how we make profit.

Sounds like the world before Descartes.

Given the right mathematical algorithms, you can go and be competitive. In the future, we won't have a choice. It does work, and it's very competitive. If you don't do it, and your competitor does, you're in trouble, and that'll force us all to change.

Is there anything that Demand Flow can't help you with?

It's mathematical algorithms for process and product, so if you have a process and a product, it'll help you. If you don't have a process or product, you just have wisdom or knowledge, then I can't help you.

How about the service industry where the primary product is knowledge?

In lot of service industries, the primary product is paper. Computer programming, it's paper. Medical health care benefits, insurance? It's all paper. In those industries Demand Flow has worked extremely well. We've done it there — it's not our true love, because we prefer manufacturing, but we do a lot in the office environment.

Where is the world going to be as it finds Demand Flow is the way to go?

We're going to move to a very competitive global economy. I think Demand Flow is going to give us tools to change the direction of market demand and product innovation as quickly as possible. If you don't do it, your competitors will do it in time. We're going to see those with Demand Flow Manufacturing become very healthy and big. On the other hand, those companies that don't adopt Demand Flow might very well end up downsizing. The future could be that simple.

Web Site © Copyright 2020, 1999 by Lionheart Publishing, Inc.
All rights reserved.


Lionheart Publishing, Inc.
2555 Cumberland Parkway, Suite 299, Atlanta, GA 30339 USA
Phone: +44 23 8110 3411 |
E-mail:
Web: www.lionheartpub.com


Web Design by Premier Web Designs
E-mail: [email protected]