WILL IT SELL IN PODUNK? HARD TO SAY
Test marketing. The phrase conjures up images of researchers swarming over towns such as Duluth, Erie, and Fresno for yearlong trials of new products--sort of the brand manager's version of the out-of-town tryout for a Broadway show. Perfected in the 1960s by the likes of Procter & Gamble Co. and Lever Brothers Co., the classic test gives companies a chance to gauge consumer and retailer reactions to a new product, then fiddle with formulas, packaging, ads, and promotions. The 12-to-18-month process can cost more than $1 million, but it's a lot better to bomb in test than in a coast-to-coast flameout.
These days, though, fewer companies are bothering to see how their fledgling products play in Peoria. Instead, consumer-goods marketers such as Lever Brothers, Colgate-Palmolive, and General Mills want newer, faster, and cheaper ways to test their offerings (table). And some are skipping test marketing altogether. R.J. Reynolds Tobacco Co., for example, went national with its Camel Wides brand in March without any testing. Robert Schmitz, director of market research for Lever USA, says he rarely tests anymore: "Today, the bias is toward immediate action."
SABOTAGE. And that's why the bias is against the traditional test. A company that spots an unfilled need in today's fast-moving market had better act quickly. Leisurely tests give rivals the time to see--and seize--the same opportunity. Worse, competitors that stumble across the product in test may rush out their own versions. Such experiences have taught Procter & Gamble, once the archetypal cautious tester, to become a much speedier player. Several years ago, for example, it began testing a ready-to-spread Duncan Hines frosting. General Mills Inc. took note and rushed out its own Betty Crocker brand, which now dominates the category. "If someone else gets there first, you just have another me-too product," says Arthur Schulze, vice-chairman of General Mills.
Worse, market tests may be increasingly unreliable, because aggressive rivals now often step up their own efforts in the test area to foul up the results. "The competition will do everything within the law to make a test tough to read," says Sheldon Roesch, general manager of the Pepsi Lipton Tea Partnership and former head of Pepsico Inc.'s new-business ventures. When Pepsi tested Mountain Dew Sport drink in Minneapolis two years ago, Quaker Oats Co.'s Gatorade counterattacked furiously with coupons and ads. Mountain Dew Sport got yanked, although Pepsi says it wasn't because of Gatorade's sabotage.
Even so, companies still need a way to gauge a product's potential. Many are turning to pretesting, or simulated testing. Researchers recruit consumers at a grocery store or mall to show them samples of a new product and its ads, or just explain a product idea. A few weeks later, they call consumers for their reaction to any samples.
Researchers then massage their findings into historical data on a new product's market to get a computer-generated sales projection. "We get within 20% of actual sales 80% of the time," says Steve Wilson, president of BASES Group, a researcher in Covington, Ky. That approaches the accuracy of regular tests--provided marketers don't fiddle with the variables and muck up the findings, says Allan Baldinger, marketing research director at the Advertising Research Foundation. Pretests often take only three months and may cost $250,000, or a fourth as much as a traditional test, says Joel Rubinson of NPD Group, a Port Washington (N.Y.) researcher. Lever USA, for one, relies almost entirely on pretests.
`JUST DO IT.' Research firms are now trying to improve their computerized products to grab more dollars from budget-strained marketers. NPD Group has a pretest program designed for drugs going from prescription-only sales to over-the-counter. An Atlanta company, Allison/Hollander Inc., is working on a software program one might call "virtual shopping": A consumer at a monitor uses a joystick to cruise down a grocery aisle, scan products, and "pick" anything up to examine pricing and labeling before deciding to buy.
Armed with pretesting results, some marketers are trying a sort of rolling rollout. Instead of testing in just one or two cities, they launch their product in perhaps 25% of the country. Scanner data available within days can give a fix on how the product is doing in real life, not just in pretest. In addition, the product is sold in a market too large for rivals to distort. Within a year or two, the new product hits all 50 states. That's how General Mills rolls out such products as its MultiGrain Cheerios. "You just do it," says Alice Sylvester, senior vice-president for media research at J. Walter Thompson Co. "You just go out and find out what happens."
As consumer-goods markets become more global--and often, more alike--some companies find that a few countries can serve as a test market for a continent or even the world. That's the thinking behind Colgate-Palmolive Co.'s "lead country" strategy. The company launches or tests a product in several countries, then follows through globally. Late last year, after some pretesting, Colgate launched its new Palmolive Optims shampoo and conditioner in the Philippines, Australia, Mexico, and Hong Kong. Now, the product is being rapidly rolled out in Europe, Asia, Latin America, and Africa.
So have the new methods made the dicey business of launching new products any safer? "The speed of testing has improved, but failure is still an issue very much on product managers' minds," says Dipak Jain, associate professor of marketing at Northwestern University's J.L. Kellogg Graduate School of Management. No wonder. Estimates of new products' failure rate vary hugely, but that rate could be anywhere from 66% to almost 90%. In a 1991 survey by Weston (Conn.) researcher Group EFO Ltd., marketers expected 86% of their new products to fail, up from 80% in 1984.
COMEBACK TIME? Some marketers even think classic test marketing may make a modest comeback. It's one thing to skip testing when you're spinning existing products into line extensions, such as Camel Wides. But if companies really want to attract more consumers, the argument goes, they'll have to introduce completely different products. And truly new products will need thorough testing.
There are signs of a new interest in testing. For example, BehaviorScan, an Information Resources Inc. division that recruits consumers and grocers to run tightly controlled tests in eight towns, reported a 17% increase in sales for the first quarter. And marketers point to Lever's experience with its Lever 2000, a combination deodorant and moisturizer soap. The brand became one of the biggest new-product successes in recent memory. But it did so only after spending two years in tryout in Atlanta--Lever's only major test in the last decade.SKIPPING THE TEST
Some alternatives to the traditional 12- to 18-month market test:
PRETESTING Show a few consumers samples of new products along with ads to gauge
COMPUTER MODELING Use historical data on similar products to turn small samples
of data into sales projections
ROLLING THE DICE Introduce a new product region by region, fixing ads and
promotions along the way to going national
FOREIGN FLING The lead country concept calls for trying out a product in an
overseas market, then rolling it out globally
Christopher Power in New York