Product planners, Lutz told me, approach a new car design as if it was a Harvard Business School case study, creating beautiful, profitable spreadsheets that don't translate well to the real world. Lutz told the following story about how it works in practice:
In the planning process for a new car, you're got a design that's kind of ready. It's going to be a new Chevy Malibu and everybody has agreed that we'll plan for 220,000 units. So you come up with an average selling price to dealers of $22,500 and you deduct the costs and have your gross margin times 220,000 vehicles. You end up with a huge blob of money.
Everybody signs off on that and the bean counters start going to work. They work with the product planners and say, "We don't have to put in 12oz carpeting, we'll put in 8oz carpeting. Why do we need this expensive protein vinyl? Let's go with regular vinyl. Why are we painting the plastic parts? Just grain the plastic. It's not going to be that much of a difference. Why do we have so many chrome moldings?"
And they all think they are heroes. They'll say, "I was on such and such a program and I was able to boost the returns by two full points." And you'll say, "But how did the car do?" And they'll say, "Not well, I think the marketing guys blew it. But man, it was a great program" because it was one that looked wonderful on paper when they got done with it.
In pinching those pennies, the car companies end up ruining the consumer experience. That, in turn, Lutz contends, forces the American automakers to offer thousands of dollars of incentives -- more than their foreign competitors -- to get people to buy the cars, which erodes the margins on the cars. But that last step, the incentives, isn't included in the spreadsheets created by the bean counters.
But, of course, the destruction of the cars' finish is never projected to impact the sales of the cars because the look and feel of the car aren't quantifiable. Lutz said that also meant that the predictions made by product planners tend to be terribly off. "They don't know anything about cars, but they are very good at macroeconomics. They generate tons of paper, but it's all useless," Lutz said. "You're far better off is to have four or five senior people who have a good feel for the market, who can look at the car, know where it's going to be priced, look at the competition, and then come up with a [sales] number." He pointed to the Pontiac Solstice, which GM's planners thought might sell 6,000 units, but which sold 45,000 in its first year, and to the PT Cruiser, which was projected to sell 5,000 units as a "retro" vehicle but which sold more than 120,000 units in 2001.
"Look at the past record. Our volume estimates are always wildly off. There are almost none that come close to what happened," Lutz said. "And yet [the product planners] say, 'That might be true but that's no reason that we shouldn't try to take a rational approach.'"