Leah Archibald: My guest Hank Marcy served as VP of Global Product Technology for Whirlpool, and as Senior VP of Product Design and Engineering at Bissell. But he really saw the potential of the APAC region when Johnson Controls asked him to move to Japan to head up a 5-billion-dollar joint venture between the HVAC company and Hitachi Air Conditioning. As VP of Global Product Development, Hank integrated design labs across 6 locations and over 1200 people.
What he learned in Japan was that a strategy of empowering local design teams and sharing data across the organization can help manufacturers manage cost and release winning products into the marketplace.
Hank Marcy, welcome to the podcast.
Hank Marcy: Thank you Leah. It’s a pleasure to be here.
Leah Archibald: I’m wondering: When you were asked to move to Japan to integrate Johnson Controls with Hitachi, did you go in with any misconceptions about working in APAC that changed once you got into the actual contact?
Hank Marcy: Sure. For example, Johnson Controls, Whirlpool, and even Bissell to a significant degree operated as a global company. We developed products for global markets and sold them globally. And we did that in a coordinated fashion by looking at needs across all the markets and then working to develop the best products to meet those needs. We made our products modular in order to meet very specific needs in specific markets.
What I found in the joint venture was that Hitachi had run the business very, very differently—almost as a collection of independent entities. The basic product development model was to develop products for Japan first, and then see if they could be sold somewhere else. This process took years from the first product concept thought to seeing if it could be sold in another country. It had not occurred to me that a company operating on a global scale would not operate in a global fashion.
Leah Archibald: Now, I assume when you went into Japan you wanted to bring them up to speed as part of the global product development organization.
Hank Marcy: Yes. In the joint venture between Johnson Controls and Hitachi, the Japanese teams had a very strong residential product unit and a very strong commercial product unit. But I also inherited about 200 engineers in China that had been working on very similar products under a different brand. And then I had a team in India that was quite strong and that we wanted to add too. So over time—over the course of about two and a half to three years—we began to take the accountability for different commercial products and divide those accountabilities between the groups in Japan and in China and in India. This forced strong collaboration. They worked as a team, meaning people in China and India reported to people in Japan and vice versa.
The idea is to be modular. You can really break down modular into two central concepts. One is the concept of the modules themselves, and two is the concept of the interfaces between the modules. If you can define the interfaces well enough and completely enough, then you can allow individual teams to work on modules and know they’re going to fit back together.
We generally thought in terms of three types of modules. 1) Modules that are very capital intensive that you don’t want to change very often at all. Structure is a good example—like the structure of a washing machine. Once I invest in the tooling to produce that structure, I probably don’t want to invest in tooling again for as long as possible. Then there are 2) modules that are very technologically intensive that might need to change because of the technology. You don’t want to change those very often either, but probably a bit more frequent than the capital-intensive modules. And then there are 3) the modules that are the features and benefits that are very tangible to the customer. Those modules may need to change more frequently, and they may also have regional differences. Maybe in this country, I want an ice and water dispenser with a fancy LCD display, but in India, I don’t want a water dispenser, but I do want the same size refrigerator. So, I need a door structure that can have an ice and water dispenser or not and decide based on the needs of the individual markets.
Leah Archibald: I imagine that if you build your products based on this modular system it’s easier to rationalize the cost of all your components.
Hank Marcy: Yeah. You really work to minimize the number of unique parts. Or maybe minimize is not quite right: you HAVE the right number of unique parts. I always use the example of water valves. A washing machine has water valves. Well, how many different water valves do you need to meet the needs of different regulations, water-types, and piping sizes around the world?
Leah Archibald: I have no idea; what’s the answer?
Hank Marcy: We counted at Whirlpool one time, and we came up with 132 different water valves.
Leah Archibald: That’s sound like a lot.
Hank Marcy: A lot. Then we said: How many could we get by with if we standardized across the world? And the answer was around 30. So now we took the volume across 132 different part numbers and put it into 30. Well, we got very significant price reductions from our vendors by increasing the volume on a part that is now more standardized. And we repeated that story across thousands and thousands of parts and pulled massive amounts of money out of the supply chain.
Leah Archibald: Now, how do you do that? Do you look at parts one by one or do you look at a grouping of parts? How do you get that cost savings out of it?
Hank Marcy: It’s difficult because most procurement organizations work at a data level that’s very high level. They know what’s a valve, but they don’t know anything else about it. Like how many inlets or outlets? What’s the diameter? What’s the voltage? At Whirlpool, the way to get the most money out was to limit the number of unique architectures at the highest level. So how many unique side by side refrigerator designs do you need? We started out with eight and I think they’re now down to four. They got there by eliminating all of that additional complexity. Maybe there were small dimensional changes on those other models that ultimately weren’t necessary in the market. Now we have a more streamlined, more cost effective, and more attractive offering to our buyers.
Leah Archibald: The focus on streamlining the component parts of all your products—is this something that the product managers need to own? Or does this have to be a strategic value that goes all the way up?
Hank Marcy: It impacts every aspect of the business, because of the complexity. Just think about inventory. Take the example of the inventory you’d need to handle your 132 valves versus 30 valves. If you had four people managing the 132 valves, you could probably manage the 30 valves with one. And so it goes: on to procurement, to manufacturing, to logistics in your factories, to your warehouses.
Anything you can do that reduces complexity makes the business simpler to run. The downside, or the pitfall is that if you go too much to standardization and lose your ability to customize for the markets, then you get into trouble. That’s why we go back to the 3-part model for modulization: rapidly changing modules for customer needs, moderately changing modules for technological needs, and rarely changing modules due to capital needs. It’s a good framework because people can understand that there are different levels of standardization, depending on what part of the product we were talking about.
Leah Archibald: How does data sharing help make this possible?
Hank Marcy: Where data sharing becomes important in the context of a global engineering organization is in design reuse. If it’s easier for an engineer to design a new part than to find the design for an old part, he is going to design a new part every time. So, you need to make it easy to find, not just to search for, but to find existing parts that can meet your needs. And this comes back to classifying your parts and describing your parts in more detail than what a typical procurement system will do.
Leah Archibald: In the big picture, manufacturers are really facing two dilemmas: How do I make it? And where do I make it? How do you think a global company should go about addressing these questions?
Hank Marcy: That’s a really good question. In manufacturing intensive companies like Whirlpool or Johnson Controls, you often find an advanced manufacturing department whose sole job is to prove out new manufacturing processes, with an eye towards greater efficiencies, faster speeds, and lower cost. That addresses how to make it.
For where to make it, I’ve been part of building logistics and factory models that can tell you the relative costs of manufacturing in different locations. Having those models is really very valuable. You don’t always have to have them at decimal point accuracy, but even if you can do relative comparisons such as: should we make it in the Middle East? Should we make it in Thailand? Should we make it in China? Should we make it in Vietnam? Those are important questions to ask that require a fairly detailed analysis to answer. Having good models of your factories and logistics makes that easier to do.
Leah Archibald: Hank Marcy, it’s been a pleasure talking to you. Thank you so much for joining me on the podcast today.
Hank Marcy: Well, thank you, Leah. This is great, I really enjoyed it as well.