Egg rolls for dinner: What a great idea! That was the thinking behind the decisions to launch a line of La Choy frozen egg rolls in 1988. Not measly little appetizer egg rolls, mind you: What managers for the Hunt-Wesson Inc. brand envisioned were big, meaty egg rolls that a consumer would happily eat as a main course. The egg roll offensive would complement another new idea, the Fresh and Lite line of low-fat frozen Chinese entr es. After all, La Choy was a well-known brand name, thanks to its canned goods; ethnic cuisine was soaring in popularity; so were frozen meals. What could go wrong?

Plenty, as La Choy discovered when it rolled out its egg rolls into the East and Midwest. For starters, they couldn't be microwaved, because the shells got soggy. And it took a very long 30 minutes to heat these giant entr es in an oven. The other Fresh and Lite products didn't set the world on fire, either. For one thing, there was the name. "It sounded more like a feminine-hygiene product," says Linda Krakowsky, an ad executive who worked on the campaign. "And it was hard to say it was fresh anyway, because it was frozen." Two years later, La Choy executives pulled the plug on both the monster egg rolls and the Fresh and Lite line: Today they say they never had the market clout to make these products succeed.

In the long and inglorious history of new-product flops, La Choy is not alone. Put in the context of such historic screwups as Ford Motor Co.'s Edsel

(estimated losses: $250 million), RCA's VideoDisc player ($500 million), and Time Inc.'s TV-Cable Week ($47 million), a few soggy egg rolls don't seem all that catastrophic. Producing flops is part of doing business in every industry, from consumer products, where relentless competition for store shelf space drives many new products to quick extinction, to electronics, where rapid technological change dooms many newcomers even after a promising start. Remember the Osborne portable computer?

IS FAILURE ESSENTIAL? And the flops and near-flops just keep coming. Among recent events: Dell Computer Corp. admits it messed up its notebook launch. Toyota Motor Corp. records disappointing initial sales for its much-heralded T100 pickup truck. Coca-Cola Co. struggles to succeed with what looked like a nifty idea--a tiny soda fountain it had designed for office use.

Overall, the new-product battleground is a scene of awful carnage. Chicago consultants Kuczmarski & Associates just studied the success rates for 11,000 new products launched by 77 manufacturing, service, and consumer-products companies. They found that only a little more than half--56%--of all products that actually get launched are still on the market five years later.

Other studies peg the long-term success rate of new products closer to 65%. But everyone agrees that most ideas never even make it into test markets. Companies had to cook up 13 new-product ideas before they came up with a winner, according to Kuczmarski. "Clearly, all is not well," says Robert G. Cooper, a professor of marketing at McMaster University in Hamilton, Ont. A specialist in new-product development, he cites a Booz Allen & Hamilton Inc. study that some 46% of all new-product development costs go to failures. That's an improvement from the 1960s. Still, he adds: "If half of a factory's output ended up as defects, you'd shut the place down."

Clearly, too, such inefficiency makes a tempting target for today's lean and hungry corporations searching for ways to cut costs and increase productivity. For U.S. companies that spent the 1980s improving efficiency and boosting quality, and that now face vicious competitive battles in slow-growth markets, raising the success rate of new products increasingly looks like one of the final frontiers. "Industry has been enormously successful at wringing operational defects out of the system," says William BonDurant, a top marketing executive at Hewlett-Packard Co. "But then you look at failure rates for new products, and they haven't changed much in 25 years. There has to be a way to drive that rate down." Adds Yoram Wind, a professor of marketing at the Wharton School: "If companies can improve their effectiveness at launching new products, they could double their bottom line. It's one of the few areas left with the greatest potential for improvement."

To be sure, there's a school of thought, ardently advocated by some executives who devote their lives to devising new products, that insists a certain rate of failure is essential. Says Brian Swette, general manager for new business at Pepsi-Cola Co.: "If you're batting a thousand with all your new products, you're doing something wrong." A perfect new-product record, he argues, means a company isn't taking necessary risks to develop new markets. Then, too, failure often lays the groundwork for a future success. The classic example: The not-very-sticky glue that technicians at 3M Co. turned into Post-It notes.

Yet that philosophical view isn't keeping companies from taking a much sterner attitude toward new-product development in a broad array of industries. The No.1 priority at the Marketing Science Institute, a research group backed by such heavies as Procter & Gamble Co. and Apple Computer Inc., is the improvement of new-product development. Companies as different as Hewlett-Packard, Motorola, Colgate-Palmolive, and Chrysler have been tackling new-product issues, too. They want to figure out how invention, in-house teamwork, and customer involvement will compress development time, unleash a flood of successful new products--and keep flops at bay.

PRODUCT CHAMPIONS. It's a riddle that requires a company to understand not just its markets and customers, but itself. Sure, says Abbie Griffin, professor of marketing and production management at the University of Chicago, product development means figuring out what customers want and developing an offering to meet that need. But, she adds, it also means considering what a company is best at, how it goes about the business of devising and marketing new products, and the path those products must follow as they move through the company's infrastructure.

And the more executives, academics, and consultants delve into what it takes to make a success, the more hurdles they discover a new product must overcome. When executives at Hewlett-Packard's Medical Products Group studied 10 of their new-product failures along with 10 of their successes, they were surprised to identify a total of 14 essential tasks that determined which products worked and which didn't. The steps covered a wide range of corporate skills. Among them: figuring out which new products play to a company's core strengths, understanding how a new product should be sold, and getting an early fix on a project's costs. "We found if you missed on just 2 of the 14 factors, you failed with your product," says Mark Halloran, chief of research and development. Other studies have pinpointed the same need to master a wide range of disciplines in order to achieve success (table).

So what are the steps to new-product nirvana? For starters, a new product must satisfy a customer's needs, not a manager's. Every new product needs a champion, of course--someone who believes in an idea and is willing to take risks to see that idea grow. But advocacy can easily turn into self-deception, dedication into wishful thinking. That's a painful lesson Steve Jobs learned. The visionary founder of Apple Computer tried to repeat his success at NeXT Inc., the startup company that developed the sleek, black NeXT desktop computer. After burning through $200 million of investment funds, NeXT stopped shipping the $10,000 computer in February of this year. It's now concentrating on its far more promising software.

What happened? As Richard A. Page, NeXT co-founder and former hardware vice-president, put it a few months ago, when he quit: "The customers know what they want." And what they don't. For starters, they did not want Jobs's optical drive instead of the usual floppy drive. The new feature made it tough to switch work from a PC to NeXT. Even after a floppy drive was added, the machine itself remained slow, and there wasn't much software available to run on it.

True, the machine had nifty features, such as hi-fi sound. But even though Jobs tried to attract various kinds of customers, NeXT never overcame its essential, customer-hostile flaws. Students found it too expensive, even after discounts. Engineers preferred desktop workstations from Sun Microsystems Inc. Although he has dropped it, Jobs insists that the NeXT computer was the right product: It was just too late to compete with the more powerful machines already out there. Yet if he had listened to customers and gone with more standard technology earlier on, analysts say, he might have had a chance.

BLIND TO THE SIGNS. Jobs was right about one thing: It's important to get to market swiftly. But it's even more important to get quality and pricing right the first time--even if that means delay. That's the painful lesson Cadillac learned with its Allant . When General Motors Corp. launched the model with great expectations in 1987, Cadillac managers had hoped that the $54,700 coupe would bestow an aura of sexy Euro-styling on the division's whole line and expand its customer base to the younger buyers being lured away by BMW and Mercedes-Benz. It didn't exactly work out that way: The Allant went out of production July 16, the victim of a too hasty launch and a failure to offer the right mix of price, quality, and features to finicky consumers.

For starters, the car, with its 170-horsepower engine, was underpowered compared with foreign rivals. The body, handcrafted at the Pininfarina workshop in Turin, Italy, was attractive, but not especially distinctive or well made. The roof leaked, and squeaks and wind noise marred the luxury-car hush.

These were all clear signs that the car's handlers should have waited and ironed out the bugs. But signs work only if they're heeded. One executive who worked on the Allant later on says that Cadillac couldn't bring itself to delay the launch. "They had made a big hoopla about the introduction of this car, and when the car wasn't ready, they didn't want to make the hard choice and hold back," he says. The result was a car too small and expensive for core Cadillac buyers, but not really good enough to lure import buyers. No wonder Cadillac sold fewer than half the expected 4,000 Allant s in the 1987 model year. And it never sold even half its goal of 7,000 cars in subsequent years.

By the time GM decided to pull the plug, Allant (now priced at $61,675) had finally become the ultrasmooth, high-performance luxury coupe it was originally intended to be. Cadillac had added the 295-horsepower multivalve

V-8 Northstar engine and an electronically controlled transmission. But buyers were thoroughly confused. It was five years too late.

Taking a different view, Cadillac's general manager, John O. Grettenberger, says: "The car was a victim of economics, not a failure." He says a financially strapped GM finally had to focus on its core models. Allant 's initial problems, he says, were not a result of haste. They were normal for a new venture in a market that Cadillac was just entering. And although he "would have liked to see it sell in greater numbers," Grettenberger says the Allant taught Cadillac valuable lessons in marketing and technology that have made the Seville Touring Sedan a success.

Of course, if Allant had gone through a more successful launch, it might have had a longer life. But even getting the new product right from the start is no guarantee of success: It must still be sold in the right way, through the right channels. A study by Cooper of McMaster shows that new-product managers double their chances of success when they successfully match a new product with the right sales force and distribution system.

"EMOTIONAL SELL." Huffy Corp, for example, the successful $700 million bike maker, did careful research before it launched a new bicycle it dubbed the Cross Sport, a combination of the sturdy mountain bike popular with teenagers and the thin-framed, nimbler racing bike. Huffy conducted two separate series of market focus groups in shopping malls across the country, where randomly selected children and adults viewed the bikes and ranked them. The bikes met with shoppers' approval. So far, so good. In the summer of 1991, Cross Sports were shipped out to mass retailers, such as the Kmart and Toys 'R' Us, chains, where Huffy already did most of its business.

That was the mistake.

As Richard L. Molen, Huffy president and chief executive, explains the company's slipup, the researchers missed one key piece of information. These special, hybrid bikes, aimed at adults and, at $159, priced 15% higher than other Huffy bikes, needed individual sales attention by the sort of knowledgeable salespeople who work only in bike specialty shops. Instead, Huffy's Cross Sports were supposed to be sold by the harried general salespeople at mass retailers such as Kmart. Result: "It was a $5 million mistake," says Molen. By 1992, the company had slashed Cross Sport production 75%, and recorded an earnings drop of 30%.

A corollary of "Know thy sales channel" is another rule: Don't sell a product just because you need something new to pump through the sales channel you have already mastered. In 1989, BIC Corp. introduced a small $5 glass flask of perfume, to be sold in the supermarkets and drugstore chains where BIC had so much distribution clout. After much hoopla, BIC sold $5.6 million worth of Parfum BIC in the U.S. before withdrawing it from stores in the first half of 1990. The cost: about $11 million.

The problem: Perfume is hardly as disposable and utilitarian as a bag of razors. Says Mark A. Laracy, president of Parfums de Coeur Ltd., in Darien, Conn., which does about $70 million a year by mass marketing knockoffs of pricey scents: "Fragrance is an emotional sell to women. But the BIC package wasn't feminine. It looked like a cigarette lighter."

TOO-SUBTLE QUAFF. Even if the product is fine and its distribution channel is right, though, it may still come to grief if consum-ers don't understand its benefits. Three years ago, for example, brewers fell in love with dry beer, quaffs that were supposed to have a cleaner finish. Many thought dry beer could be the boost their industry needed. Anheuser-Busch Cos., Coors Co., and a clutch of smaller foreign brewers shipped 4.6 million barrels of dry beer in 1990.

But by 1992, shipments had fallen to 3.7 million barrels--a minuscule drop of the industry's 197 million-barrel total--and brewers will be lucky to do even that much this year. Pretty sorry returns on the $40 million-plus spent in advertising for the category. Coors Vice-Chairman Peter H. Coors figures dry beer appeals to only 2% of all beer drinkers. "At the time, it seemed like an intelligent decision," he says. "Knowing what we do now, we probably wouldn't have gone with it."

What Coors knows now is that consumers, who usually only spend a few seconds in a store or bar choosing a beer--or any product--are not going to take the time needed to understand a too-subtle product like dry beer. As marketing consultant Jack Trout puts it, "Nobody can figure out what the hell dry beer is. The opposite of wet beer? It's never been explained." And it probably never can be.

Numerous as the pitfalls are, it's possible for companies to avoid them. But improving the odds in new-product development often takes some fundamental rethinking of the way a company approaches its markets and manages its own operations. Chrysler in the 1980s, for example, was a textbook case of how not to devise new products. Ignoring the need for continuing innovation, the carmaker relied throughout the decade on the K-car platform it introduced in 1981. Such models as a too-narrow New Yorker sedan were launched and flopped. In addition, Chairman Lee A. Iacocca used his influence the wrong way, walking into Chrysler's styling studios and ordering up more chrome or opera windows with a wave of the hand.

As sales flagged in 1989, Chrysler President Robert A. Lutz finally convinced top executives that the company had to change its ways. So Lutz and Fran ois J. Castaing, Chrysler's engineering chief, threw out their old, compartmentalized approach to new-product development, in which a project would pass from research to engineering to manufacturing, and finally on to marketing. Instead borrowing from Japanese practices, cross-functional teams of engineers, market researchers, marketers, stylists, and manufacturing engineers began working together to design and build new models. The team approach has chopped by 40% the time it takes to develop a new car or truck. One important result: the hot LH line of sedans.

SIX STAGES. Other companies are also using team approaches to hone their new-product processes. Colgate-Palmolive Co. had a string of new-product flops in the 1980s, most notably Fab 1 Shot. The product, a slug of detergent, softener and an antistatic ingredient in a soluble bag, was supposed to save consumers time and hassle, but instead it irritated users, who couldn't adjust the mix to suit their own needs. Since 1989, Colgate has overhauled its new-product process to concentrate cross-functional teams on far fewer product ideas and speed up global rollouts of the most promising products.

Now, only about 20% of ideas make it to prototype, down from around 50%. And each idea must meet specific criteria at six different stages that lead from development to achieving a commercially viable product. While Colgate hasn't yet won a reputation as a new-products runaway winner, its latest introductions, such as Precision toothbrush and Stand-up toothpaste, have been successes.

Other companies have learned to get back to their core strengths. Campbell Soup Co. dismayed consumers and stockholders alike in the 1980s by pouring out almost 160 new products a year, with a success rate of at most 20%. "We didn't stay close to home," says former North and South American Div. President Herb Baum. The failures included such notable bobbles as Campbell's Fresh Chef line of fresh salads and soups, which had a shelf life of a week. The company was constantly misjudging which dishes would sell and which wouldn't. "We never knew what to make," says Baum. The result was spoiled food and irritated supermarket owners.

Now Campbell has cut back new products 20%, to around 120 a year. And those are mostly in areas it knows best, such as canned soups, sauces, and baked goods. When the company ventures into new areas, it is with something such as Pepperidge Farm canned gravy, which uses technologies Campbell understands and brand names that it has already established.

INSPIRING CASES. The likes of Chrysler, Colgate, and Campbell have plenty of incentive to improve the odds for their new offerings: There appears to be a strong correlation between new-product success and a company's health. According to Thomas P. Hustad, editor of the Journal of Product Innovation Management, the companies that lead their industries in profitability and sales growth get 49% of their revenues from products developed in the past five years. The least successful get only 11% of sales from new products.

There are also plenty of specific cases to inspire and terrify. Gillette Co.'s stock has soared since it introduced its new Sensor razor. IBM's has floundered since it amply demonstrated its ineptitude at creating new product lines that pay. But IBM has also shown how it is possible to change: Its small ThinkPad computer is a big new-product success.

Whether it's frozen egg rolls or microchips, the message is clear: A company cannot avoid every flop. But if it learns from its mistakes, it can surely flop a lot less often. And understanding failure is clearly a key to success in this ferociously competitive decade.

      How a company can improve its success rate in product launches
      1 ASK YOUR CUSTOMERS Don't launch a product just because the engineering 
      department loves a new technology. Consult users at every step, from idea stage 
      to rollout.
      2 SET REALISTIC GOALS  A new product might be sure to produce $20 million in 
      sales. So don't make it a loser by aiming for $40 million.
      3 BREAK DOWN WALLS Passing off a new product from one department to the next 
      risks potentially disastrous foulups. Instead, have research, marketing, and 
      manufacturing work together from the start.
      4 CREATE GATEWAYS Don't let a new product gather dangerous momentum. At each 
      stage of a product's development, make sure it meets specific criteria of 
      manufacturing viability, customer acceptance, sales support and budget planning.
      5 WATCH THOSE TESTS  A test market may succeed just  because consumers are 
      sampling a 
      new product out of curiosity. Don't get carried away by initial results. Test 
      long enough to get a real sense of a product's potential.
      6 DO YOUR POST-MORTEMS Managers tend to run away from their flops. Don't. 
      Formally review what went wrong and apply those lessons to the next launch. 
      Reward managers who learn from mistakes.
         FORD'S EDSEL It had innovations galore--and quality problems from stuck 
      hoods to defective power steering. Estimated loss per car--almost $1,117, or 
      $250 million.
         DUPONT'S CORFAM A synthetic leather supposed to do for shoes what nylon did 
      for stockings. Leather was just better. Cost: $80 million-$100 million.
         POLAROID'S POLAVISION Edwin Land used Polaroid's wet chemistry technology to 
      develop an instant movie camera. But videotape technology was far better.
         UNITED ARTISTS' HEAVEN'S GATE Almost $30 million over budget, this western 
      movie bombed so badly it almost destroyed UA.
         RCA'S VIDEODISC Supposed to capture the video recorder market. A tiny 
      problem: It couldn't tape television shows. Loss: $500 million.
         TIME'S TV-CABLE WEEK A  bid to outflank TV Guide. Cause of death: Ballooning 
      costs to customize editions for each cable system. Loss: $47 million.
         IBM'S PCjr The awkward Chiclet keyboard. The slow microprocessor. The 
      unattractive price. The late launch. IBM at its worst. Marketing cost: $40 
         NEW COKE Coca-Cola's answer to Pepsi's sweeter formula. Provoked a national 
      uproar from old-formula loyalists. Watch for new formulation, Coke II.
         R.J. REYNOLDS' PREMIER This "cigarette" didn't burn. It didn't emit smoke. 
      It didn't taste good. Its failure persuaded CEO Ross Johnson to launch his 
      equally disastrous buyout attempt.
         NUTRASWEET'S SIMPLESSE The fat substitute that would change the way we eat. 
      But the market is swamped with substitutes, and many consumers like fat.
Before it's here, it's on the Bloomberg Terminal.