Skip navigation
Market-testing new products earns A-plus from national chains

Market-testing new products earns A-plus from national chains

When Denny’s was developing its new Allnighter menu, consumers were not just invited into the chain’s test kitchen to help the culinary team develop late-night dishes, they were brought in under the cloak of darkness.

“We did it at 11 o’clock at night,” says Mark Chmiel, chief marketing and innovation officer for the 1,500-unit family-dining chain. “It had the vibe of what late night is all about.”

As the restaurant industry grows more competitive, restaurant operators are becoming increasingly savvy in conducting the tests that will determine whether new products and promotions are rolled out systemwide or shelved.

From the more thorough vetting of ideas before they leave the corporate kitchen, to the more careful selection of test sites and the use of numerous benchmarking hurdles, test-marketing strategies have grown in scope and sophistication.

While test-marketing methods vary widely from company to company, restaurant executives agree that testing is a necessity that should be sidestepped only when the risks involving a new-product launch are extremely minimal.

Not all restaurant chains involve consumers so intimately in product development and testing as Denny’s, but officials at the Spartanburg, S.C.-based chain have discovered such involvement is essential, especially for breakfast items.

“The No. 1 thing they’re looking for is new-product ideas,” Chmiel says. “They’re looking for variety, they’re looking for choice.”

And what they’re looking for in the morning is a “real breakfast,” something Chmiel learned—somewhat to his surprise—when Denny’s tested its to-go breakfast program and domed carry-out packaging.

The chain included handheld breakfast sandwiches in the test, but consumer reaction essentially was, “‘I can already get a sandwich from a QSR,’” Chmiel says.

“That was a major thing we learned in test market,” he says. “We decided to hold off with sandwiches and go with a complete breakfast.”

Sales of complete to-go breakfasts, which come inside a domed package to keep the items warm, have “exceeded projections,” Chmiel says.

Denny’s usually tests new products for four to six weeks and supports the test with the same live-media mix it would use for a systemwide launch of the product.

But Denny’s scrapped live media when it tested Sizzlin’ Breakfast Skillets in the spring of 2007 because it was a product no one else had, and “we did not want competitors to thwart us,” Chmiel says.

In that case, the skillets were tested in Boise, Idaho, because a franchisee there was a strong supporter of the product, Chmiel says. So strong, that the franchisee sent e-mails to other franchisees systemwide urging them to get behind the product for a national rollout.

“By gaining support of a key franchisee we were able to make that happen,” Chmiel says.

Sizzlin’ Breakfast Skillets rolled out systemwide in August.

Where a restaurant tests a product depends on the nature of the test, says Aric Nissen, vice president of brand marketing for Edina, Minn.-based Dairy Queen.

If a chain is trying to predict national demand for a product, the test covers multiple regions. To test a new-product platform that will receive national marketing support it’s best to choose a concentrated area, Nissen says.

Dairy Queen, which has nearly 5,000 U.S. units, can test a new Blizzard flavor in 20 stores, backed with a point-of-purchase campaign, and predict how it will perform nationally, he says.

The chain’s Girl Scouts Thin Mint Cookie Blizzard was tested in stores across the country and was supported with point-of-purchase material. Other products, like the Waffle Bowl Sundae, may get radio and print support and some TV ads during test, Nissen says. A simple flavor test might last only a month, but testing a new-product platform “takes an extended period of time” to measure the product’s durability, customer repurchase intent and how the new product is affecting sales of established products, he says.

Like all large restaurant chains, Dairy Queen is “continuously testing” new products, Nissen says. The chain employs a three-year time frame for tests, meaning in any given year it will be rolling out a product that was tested the previous year, testing a new product for possible rollout the next year, and coming up with ideas for tests two years out.

Test-marketing has increased since the early 2000s, when chains were more interested in value menus, Nissen says. But since the economy has gone bad there’s a new emphasis on product and flavor differentiation.

Dairy Queen has “tested everything” for five or six years, but that will change in 2009. Every new flavor introduction may not be fully market-tested, but all of them will be fully researched, he says. For example, members of Dairy Queen’s Blizzard Fan Club “may want more variety than we can test in a given year,” Nissen says, leading the chain to launch a new Blizzard flavor without a full test.

“It’s a question of whether you can take small risks,” Nissen says. “Small risks are acceptable.”

Despite the “large expenses” in testing, every test can be viewed as a success, even if consumers rejected the product, he says. Rolling out an untested product that fails in the marketplace is more expensive than conducting a test and deciding not to launch, Nissen explains.

Boston-based Au Bon Pain, which has more than 200 units, has done so many tests that it has a complete set of benchmarks to determine whether a test product would succeed in a systemwide rollout, says Ed Frechette, senior vice president of marketing.

“Lunch, salads—name a category and we have a benchmark for it,” he says.

The chain tested 40 products last year. Tests last six weeks in six cafes that are representative of the Au Bon Pain system and are conducted in such cities as Chicago, Boston, New York and Washington, D.C.

But the normal procedure had to be altered when the chain began testing is Portions menu, 14 snack-sized items each containing 200 calories or fewer.

“We started with one cafe,” says executive chef Thomas John. “We had no history testing a concept like this. This is the first time we had so many items at once.”

The chain started out with 23 different items but narrowed the line to 14. At one point almost 40 cafes were involved in the test before the menu was rolled out earlier this year. The Portions menu now accounts for 2 percent of overall cafe sales, Frechette says. Even with the chain’s benchmarks predicting how test products will fare in a systemwide rollout, testing involves intangibles, “and Portions is a great example,” he says.

Consumers liked all 23 items, but Au Bon Pain executives had to determine the operational impact on the cafes. The final menu was narrowed to the current 14 because the larger number was too much to handle in terms of prep time, Frechette says.

El Pollo Loco has a “long, rich history of product tests,” says vice president of marketing Mark Hardison, and, like Denny’s, the chain is involving consumers earlier in the test phase.

Before a product goes into the test kitchen “we make sure we have consumer acceptance,” he explains. “We might put 20 concepts to consumers on an Internet panel and see which has the most interest.”

The test-market procedure for El Pollo Loco, which is based in Costa Mesa, Calif., and has more than 400 units, can start with an in-store test in a single restaurant and then expand to others. Tests last about eight weeks and involve eight to 12 restaurants in the expanded phase.

Test restaurants are chosen based on demographics, customer volume and such available facilities as a drive-thru, Hardison says. The restaurants include those with a higher Hispanic population, a mixed demographic and perhaps those with a heavy Anglo customer base to project test results to the entire system.

The product is refined in the test kitchen, and that’s what El Pollo Loco tests in the expanded phase, Hardison says.

Crunchy and soft tacos, the Queso Crunch Burrito, and the chain’s value menu all recently made it through the rigorous test process.

Consumer acceptance is the most important factor, but the product still has to contribute to sales.

“We have had products that had great consumer acceptance, but they didn’t add anything to incremental sales, and those products don’t get rolled out,” Hardison says.

Crunchy and soft tacos were introduced in late summer 2007 and are a “great example of a product that contributes to guest checks,” he says. “Our transactions for snacks and late night have gone up with tacos.”

The Queso Crunch Burrito made it into test markets after Internet research with consumers and was rolled out in February.

El Pollo Loco tests promotional concepts as well as products, and its value menu was a tough sell when it was tested in-store, Hardison says.

During the first phase of online testing, results were “very positive.” To take advantage of dollar deals, customers first had to purchase a combo meal, which is not a common practice in restaurants, Hardison admits. The concept failed once the menu was tested in-store because there was “a lot of confusion about having to make a combo purchase first,” he says. “It contributed to guest confusion and lower speed of service and cashiers apologizing. It was not rolled out.”

El Pollo Loco reprinted all the supporting point-of-purchase material and tested it again without the requisite combo purchase.

“It was a phenomenal success,” Hardison says.

The initial in-store failure of the value menu illustrates the obvious reason that tests are conducted in the first place—chains better not roll out a loser—and that’s especially true when a big marketing push supports a new-product launch.

“When you’re spending $3 million on media, [the product] better be well-tested,” says Mike Branigan, vice president of marketing for Culver City, Calif.-based Sizzler.

Most of the chain’s media-driven test marketing is done in Sacramento, Calif., because it’s an affordable media market, he says. Otherwise, products in test markets are supported with in-store material.

Sizzler, which has more than 200 U.S. units, tests products five or six times a year.

“We will never bring a product to market or to a store if it doesn’t hit a certain scoring platform,” Branigan says. “We have a matrix that it has to hit before going into store. Otherwise we’re kidding ourselves.”

Even products added to the chain’s popular salad buffet, which 65 percent of customers use, are tested to make sure the flavor profiles are right, he says.

There also are some products that chain executives would like to add to the menu that just don’t pass muster with consumers.

“Buffalo shrimp we loved internally but couldn’t get it out of focus groups into test,” Branigan says. “King crab legs tested from Puerto Rico to New York to Texas, and we couldn’t get it done. You’re not going to have 100 percent wins all the time.”

Going to test markets is important even for small chains, especially when they’re intent on keeping the limited menu that brought them success.

The 22-unit Foster’s Grille recently added two salads, Asian Chicken Salad and Nancy’s Garden Salad, after testing them in restaurants near corporate headquarters in Haymarket, Va.

“We never shut our doors to what’s changing in the economy and what the consumers are asking for,” says Mike Cerny, vice president of operations, but the chain does not want to make the menu too complex.

“We want to keep it limited to 12 or 14 items,” he says, which makes test-marketing even more critical in knowing where to draw the line on menu expansion.

What Foster’s Grille wants from its test-marketing is knowing what will benefit the concept and the customer base of families and kids, as well as increasing traffic and generating a return on investment.

The test phase includes comment cards for customer feedback and developing customer relationships, Cerny says.

Tests were conducted in four company-owned stores because that gave executives a “very hands-on” approach to see how customers reacted to the products, he says.

“Feedback from consumers is dictating what we do,” Cerny says.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.