But about that same time, Vince Barabba, general manager of GM's strategic decision center and the company's market research guru, convinced his colleagues to toss out conventional thinking and view the business through a big set of toy building blocks.
The blocks were Mr. Barabba's way of showing in three dimensions how his research into customer needs could help solve some of GM's biggest problems: How to wisely allocate capital spending, produce vehicles that not only satisfy but delight customers, and make a whole bunch of money.
If Mr. Barabba's theories prove themselves in the next few years, he will be partly responsible for saving the Oldsmobile division and erasing GM's embarrassing stumbles in small and mid-sized cars, minivans and sport-utility vehicles.
His theories are designed to help GM be more agile and responsive to consumers-an objective that industry observers applaud.
"It's putting the customer first rather than the sheetmetal," says Jamie Power, partner and director of the Detroit office of J.D. Power & Associates, an auto market researcher.
The key is asking prospective customers the right questions during product clinics, Mr. Barabba says.
Until his research into customer needs-GM calls it needs-based research-GM and other automakers that use focus groups were asking people to voice their opinions about vehicles already on the market. The questions limited answers to established models and market segments, and gave little or no insight into new vehicle niches.
"It's just like asking, `Do you want chicken fricassee or hog jowls?' I guess I'd pick chicken fricassee, but I'd ask if you had something else," says Mr. Barabba, who joined GM in 1985.
About 1988, Mr. Barabba and his staff began working with Applied Decision Analysts, a marketing analysis consultancy, to perfect "conjoint measurement"-a way to estimate the relative preferences people have for different vehicle attributes. The process addresses questions such as how much fuel economy would a customer trade for better acceleration, or how much would a customer pay to get more cargo space.
After more than a million of these interviews, Mr. Barabba and his team could see GM would be in deep trouble if it kept doing business as usual.
GM was doing well in trucks and the large and luxury-car segments but wasn't winning customers who buy mid-size and small cars. From the research, Mr. Barabba's team plotted the future decline of the big-car segments and the growth of small and mid-sized cars.
These findings meant that GM, in blunt terms, was making great cars for old people who'd be moving on out of the market.
But Mr. Barabba's research also revealed growth niches GM could exploit if it applied what it was learning about customer needs.
"With what GM is doing, it can assess what the needs and desires are five to 10 years from now," says Chris Cedergren, senior VP of AutoPacific Group, an automotive marketing and consulting company.
GM engineers and planners who had to use the needs research had difficulty "translating" the paper reports into visions of future market segments, Mr. Barabba says.
So, his team constructed a model built from Lego Systems blocks of the auto market that stressed consumer needs as well as traditional factors of vehicle size, segment volume and buyer demographics.
By assigning values to different colors of Legos, the team constructed a model of the market as it existed in 1991 and how it would look in 2002, with flags designating the leading vehicle in each segment. The model was 15 feet wide, 40 feet long and 4 feet tall.
From the model came a plan. GM would focus the images of its marketing divisions and develop cars that fit those images while meeting the needs expressed by customers in the niches where the divisions compete.
Looking at the product portfolio is like flying across the country at 40,000 feet to get the big picture, Mr. Barabba says. When you actually get to the time when you need to define a car's characteristics, you zoom down to 5,000 feet to get a read on the details.
"Things get a lot more clear and you find out what the tradeoffs are and you start making modifications based on customer requirements based on that portfolio decision," Mr. Barabba says.
Since late 1992, Mr. Barabba's research has figured significantly in GM's engineering and design.
Although new cars such as the Chevy Lumina/Monte Carlo and the Chevy Cavalier and Pontiac Sunfire, all new for 1995, benefited from the research, Mr. Barabba says the first true tests will be the next generation of minivans and all-new Pontiac Grand Prix, both due midway through the 1996 model year.
"If you compare the new Lumina with the new Grand Prix [both mid-sized cars], you get a pretty good idea of how you might meet different customer needs," Mr. Barabba says (see story on Page S-8).
"In the case of the Lumina, people are interested in the aspects of the vehicle that provide functionality for more than the driver," he adds.
Despite GM's advances in research, some industry observers say GM runs the risk of over-fragmenting the market.
Susan Jacobs of Jacobs & Associates, a market forecasting and planning company, says needs-based market research will keep GM in the ballgame but won't be enough by itself to assure victory.
"The approach leads to more differentiated product that is better focused on consumer needs," Ms. Jacobs says. But, she adds, "in a crowded market, people don't have time to look at all the minor dimensions of a vehicle. I think GM is going to be held back by the fact that they have so many marketing divisions."
The stakes are high for GM. Following the plan means a reinvigorated Olds-smobile, higher sales and market share and prestige regained for GM.
All thanks to a vision and one of the world's biggest sets of Lego.
Vince Barabba believes in "conjoint measurement"-GM's estimate of relative preferences buyers have for different vehicle attributes.