MARKETING & MEDIA
ADVERTISING/ By James R. Hagerty
Home-Improvement Chain Puts Science to Work on Its Marketing
Many advertising decisions boil down to someone’s hunch about what might work best. Lowe’s, the nation’s second-largest home-improvement retailer, says it has found a way to eliminate some of that guesswork.
The North Wilkesboro, N.C., company “slaughtered some sacred cows” and over-turned a few long-standing ad practices as a result of tests it ran last year, says Dale Pond, executive vice president, merchandising and marketing. For instance, Lowe’s recently banished cartoon figures from its advertising circulars and is putting more emphasis on TV spots.
“We’ve got to be much smarter in the way we use our dollars,” Mr. Pond says. That is particularly vital as Lowe’s spreads out from its base in the Southeast to California and other parts of the country where it is little known. Lowe’s yearns for the kind of household-name status enjoyed by its much larger rival, Home Depot Inc.
The tests were run with help from QualPro, a Knoxville, Tenn., consulting firm, which touts a complicated statistical tool called multivariable testing, or MVT. The technique allows companies to test many changes in procedure at once, rather than trying them out one at a time. QualPro has used that approach to help chemical makers refine production processes, marketers sharpen direct-mail pitches, and the National Enquirer tweak its front page.
With the help of QualPro, Lowe’s decided to experiment with the cover pages on its advertising circulars, which are stuffed into newspapers or otherwise delivered to between 30 million and 40 million homes about 25 times a year. Mr. Pond says he encouraged colleagues to propose all kinds of changes, even “crazy” ones. The company settled on 29 changes to test; those variables included putting fewer items on the cover and using photographs instead of cartoons.
According to QualPro’s mathematicians, 29 variables gives rise to about 537 million different combinations of those variables. No one could test all those combinations. So QualPro used advanced geometry to select
36 combinations likely to demonstrate which of the changes would make the biggest impact, says Art Hammer, a principal at QualPro. Mr. Hammer says the technique is analogous to polling a small, carefully chosen sample of voters to predict how millions will vote.
Lowe’s tested those selected combinations by preparing 36 different versions of its cover and trying them out in different markets. The company then judged which changes were helpful by examining sales data in various markets. Mr. Pond says Lowe’s concluded that it works better to put three or four products on the cover than to squeeze in eight or 10. That lesson helped silence Lowe’s merchandisers who used to clamor for displaying as many items as possible.
Lowe’s also found that pictures of real people are more effective than the cartoon
figures it was using. Many customers apparently prefer that Lowe’s present itself as authoritative rather than amusing, Mr. Pond says. (Nonetheless, a spokesman for Home Depot says that chain is standing by its cartoon mascot, Homer D. Poe.)
In addition, Lowe’s looked at the size of its circulars. Beyond a certain number of pages, Lowe’s discovered, additional pages don’t generate more sales. So Lowe’s is reducing the maximum page count to around 48 from 72.
Other tests measured varying doses of television spots, newspaper ads and sports sponsorship. “We learned that TV has an even more important role than we thought,” Mr. Pond says. Partly as a result, he says, the company’s TV-ad spending will rise more than 20% this year. The tests also showed that sports marketing creates a strong bond with customers, Mr. Pond
(over please)
Home-Improvement Chain Puts Science to Work on Its Marketing
says. That finding affirmed the company’s strategy of sponsoring Nascar auto racing.
Lowe’s says the tests were unrelated to its recent decision to put its $100 million broadcast-ad account into review. The incumbent agency, W.B. Doner, is competing with five other finalists for the work.
Running the tests “isn’t a cheap exercise,” says Mr. Pond, who declines to provide cost details. But he thinks Lowe’s managed to measure things that he once thought either couldn’t be measured or would be too expensive to measure. Lowe’s will save millions of dollars by eliminating practices found to have little or no effect on sales, Mr. Pond says.
Not everyone in the company was so sure of the tests’ validity, concedes Aubrey Junker, manager of business-process improvement at Lowe’s, who goaded Mr. Pond into giving MVT a try after attending a QualPro seminar. “We still have skeptics and probably always will,” he says.
Indeed, such tests are “subject to all kinds of contamination” from other things affecting sales, such as weather, notes Frank Mulhern, an assistant professor of marketing at Northwestern University. (QualPro says its methods take such factors into account.) But Mr. Mulhern and others say properly designed tests can be good investments. Even if the tests aren’t 100% reliable, says George Belch, a marketing professor at San Diego State University, they are better than “shooting in the dark.”
