Shrinking Product Development Time

A few years ago, if you asked an auto maker about shrinking the amount of time it takes to bring a new product to market, you were treated to a demonstration of dazzling new digital technologies. There were tours of virtual-reality studios that allow computer-generated designs of new cars and trucks to be viewed life-size and in 3-D with the aid of special glasses.

They would take you to a darkened room where you could “sit” in a virtual computer-generated interior that was projected in 3-D around you, demonstrating a designer's ability to interact and evaluate a new interior without building a physical prototype.

Some auto makers, such as Ford Motor Co., showed off 50% scale holographic images of new vehicles and expressed hope that someday designers would be able to get rid of the bulky 3-D glasses and other hardware and visualize 3-D representations of mathematical models in holographic form.

It was all about eliminating the time, effort and cost involved with creating physical prototypes. And it didn't hurt that much of the technology was fascinating, fun and — for once — something just about anybody could understand. In a field laden with confusing jargon and acronyms like CAD/CAM (computer-aided design/computer-aided manufacturing), PDM (product data management) and PLM (product lifecycle management), VR (virtual reality) suddenly gave design engineers something to talk about at cocktail parties.

But today, as auto makers in many cases have chopped the time it takes to develop a new vehicle from 36 months to 24 — and only 16 or 14 months for specialty vehicles like the Hummer H2 and Ford GT, respectively — they are taking a decidedly different approach to speeding the process further.

In a surprising trend revealed by a recent study by the University of Michigan's Center for Automotive Research, in Ann Arbor, MI, people and processes, rather than technology, once again are taking center stage in product development.

“It is certainly the case that respondents were more concerned with organizational and human resource approaches to improving high-level design, than with a software-tool oriented technological approach,” says the new study: Process Efficiency and Design Rework at Early Stages of the Vehicle Product Life Cycle.

“At several instances during each interview we made a point of pushing the discussion in the direction of tools. However, the respondents continually came back to organizational themes: layers of management; and, second, populating these new systems with a relatively small number of very experienced people. There was a clear sense that personal expertise was vital and that the key to success was bringing that expertise to bear in an effective manner,” writes David J. Andrea, co-author of the study.

The digital tools still are being used — along with the funny glasses — but now, as auto makers and suppliers strive to reduce product development times to almost unheard of speed — as little as 12 months from design freeze to production — they talk about people, expertise, and making sure all those mathematical models and virtual prototypes can be used throughout the corporation.

Meanwhile, major suppliers of the computer-engineering platforms for these digital tools, such as EDS PLM Solutions Inc. and Dassault Systemes SA's CATIA, are pursuing a two-pronged strategy to both enhance the value of their systems to big, sophisticated clients such as the OEMs while also moving their products down the supply chain, where the benefits of using 3-D mathematical models still aren't being realized.

“The trend is more about integration. Looking at the overall process — rather than before looking at pieces of technology that acted like a personal performance improvement — they're looking at work group, enterprise and extended enterprise productivity,” says Stuart McCutcheon, president of the Automotive Business Solutions unit of EDS PLM Solutions.

“In order to get that level of integration and transparency of data flow between your different systems, it needs to be taken for granted,” McCutcheon says. “So you'll see our tool sets really tightly interwoven and the flow of data between them being a lot higher than in the past. We're really focused on maximizing the customer's processes as opposed to five or six years ago (when) it was more of a technology perspective.”

And when engineers and designers do talk about digital prototypes and finished computer models of vehicles and components, it's usually about figuring out ways to use and re-use them better: for future models and other vehicles based on similar platforms, the same way they would want to re-use physical parts.

“If a car's made out of 200 major components and (requires) 200 major processes to put it together; you multiply those two together, that's 40,000 things that you have to do absolutely perfectly from scratch: Your FMEAs (Failure Mode and Effects Analysis), your DVs (design verification), your PVs (production verification), your PFMEAs (Process Failure Mode and Effects Analysis,” says Chris Theodore, vice president-Advanced Product Creation, Ford Motor Co.

“The research and the effort required to do all those perfectly is huge. If you can cut that in half, if you re-use 60% of the processes and 60% of the components … you get huge improvements. And therefore, you can do a better job because your probability of making a mistake is reduced. It's a big lever,” says Theodore.

Bernard Charles, CEO of Dassault Systemes, agrees that helping auto makers re-use their digital models is a top priority.

“I think the digital world (now) is helping the automotive industry specifically to create and re-use assets in such a way that they can create products while still managing production cost to make the physical product,” he says.

“People will stop always starting from scratch. Because when you start a new program, people start from scratch. You want to re-use the engineering ideas. The economy will not necessarily be the number of parts, re-using the exact same door…it's going to be about the engineering concept. Sometimes it's not the cost, it's about re-using the engineering concept,” says Charles.

If you haven't heard or read a lot about product cycles or product development times in the past few years, there's a reason for it: Auto makers — particularly Detroit's Big Three — have improved so much it's not the issue it once was. Accordingly, critics have moved on to other issues such as under-funded pension funds and shrinking market share.

After getting beaten to the punch in key product segments year after year by Toyota Motor Corp. and Honda Motor Co. Ltd., the U.S. Big Three — especially General Motors Corp. — have stepped up dramatically to chop the amount of time it takes to launch new products and react to changes in the marketplace.

Experts say Toyota and Honda still are probably the fastest — even while their specific product development processes are very different from each other — but the Big Three now are much more competitive. They're still getting beaten to the punch, the hot new cross/utility vehicle market comes to mind, but the Big Three aren't being left in the dust like they once were.

Most of GM's new product programs for instance, are on 24-month schedules, says Rick Spina, executive director, program management — planning, GM North America.

“We're continuing to cut it down, but 24 months now is pretty much our norm. We have some programs in development now at 18 months. We know we can do it, but guess what, the world doesn't stand still,” he adds.

That's incredible progress. In the 1950s GM could develop brand new cars like the Chevrolet Bel Air in just 22 months, but when government safety and emissions standards were enacted in the 1960s, bureaucracies within the companies ballooned and by the 1970s and early 1980s it took 60 months or more to bring a new car to market after everyone signed off on the final design. That spelled disaster when fuel prices soared and Americans suddenly developed a taste for smaller, front-wheel-drive cars.

That led to a dramatic and well-publicized push to so-called simultaneous engineering, where designers, engineers, and manufacturing folks all were supposed to start working together during a vehicle's initial design to speed the development process.

When GM was able to bring out the 1988 Chevrolet Corsica in just 45 months, engineers celebrated.

By the mid-1990s everyone was talking about 36-month development times, and that's when auto makers started turning to eye-popping virtual-reality tools.

Even though GM continued to improve the process further, it still was being chastised in the late 1990s for being slow and unable to react swiftly to market changes or new products like Ford's amazingly successful Lincoln Navigator fullsize SUV.

Most critics agree GM's answer, the first-generation Cadillac Escalade, was a slow and weakly designed response that cost the auto maker tens of millions in lost profits while Ford — not exactly a speedster at product development — dominated a new luxury segment for more than two years.

Spina says GM is reacting faster now. When the third-row seat on Chrysler Group's Dodge Durango was a surprise hit, GM was able to spit out a long-wheelbase version of its GMT 360 midsize SUVs with a third-row seat in record time, he says.

And GM now is bringing “better mousetraps” to market like the Chevrolet Avalanche's innovative Midgate design first, rather than last, Spina says.

But Spina — like apparently everyone else in the industry — isn't looking for more fancy virtual reality tools to further shrink GM's product-development time.

“Frankly in the last few years we probably haven't spent a lot of dollars and mental energy on buying more of that stuff. Now it's ‘How do we use it most efficiently with speed in mind?’” he says.

Virtual-reality tools “are used constantly, but if you came back today vs. what we showed you the last time (three years ago), you wouldn't see 50% more or anything. You would see us getting a lot smarter in terms of how often we use them, what we use them for, who uses them and how we make decisions from them,” he says.

During the last 12 to 18 months more of GM's focus has shifted to people, Spina says.

“How we use the tools, everything from training to decision-making to leveraging the tools to get off hardware properties. In developing our vehicles, we go through these learning cycles. I want this suspension to feel a certain way, or I want it to handle a certain payload.

If you went back a couple of generations, you would have done that by iterating hardware: I build a prototype vehicle and I drive it and play with shocks and springs, all that stuff. You go through these learning cycles. We dramatically reduced the number of hardware learning cycles because the math models enable us to predict what the hardware is going to do.”

The next big jumps in speed will be achieved by having small groups of people — rather than a hundred separate “experts” — very clearly define what the new product is and what it has to do in the marketplace. These people will then provide concrete answers to the rest of the product team when they hit the ground running, Spina says. That and being able to store and re-use the mathematical models and learning cycles on other products is moving the industry to yet another level.

“Our philosophy for the future is you will take a car and you will morph it like clay and then you will morph sub-systems. I see some of our customers saying ‘I don't want to start one-car program. I want to start five simultaneous car programs that will be engineered simultaneously from the same concepts.’ But at the end of the day you will not believe it because they do not look alike, one is a coupe, one is a light truck, but the engineering concept of the approach is the same,” says Dassault Systemes' Charles.

Charles also sees much room for improvement in the supply chain: “Technical people now, some of them in many companies are working with such basic tools and low-level processes, we need to have them grow to a new stage. That's the reality. In the end, the car isn't going to be done by one guy, it's going to be done by the power of the teams connected to each other being able to communicate efficiently,” he says.

“Ten years ago Boeing delivered the 777 (the first major airplane created without building physical prototypes). Why, 10 years later, do you still have companies that can't even check the fit of parts (on computer)? Why are we still in that situation?”

In 20 years or so, development times have gone from 60 months to 36 to 24 and now to 18, with Toyota reportedly targeting 12 months. Will it ever end?

“All the competitors are trying to figure out how to get to market faster,” says Ford's Theodore. “It doesn't matter whether it's the auto business or the cell-phone business. There's a strategic advantage to getting quicker and that has to do with the product-creation process.”

“The only reason I'd ever tell you we don't want to get any faster is because the rest of the world isn't getting any faster,” says GM's Spina. “Unfortunately that's not true. This is going to be a continuous evolution. We don't know where the bottom is.”
— with Eric Mayne