TESTING Tougher tryouts

Nov 01, 2000 10:30 PM  By

Mailers are making list tests more highly targeted Finding the right prospects for your catalog is no easy task. So many catalogers are using larger tests and sharper, more focused testing tactics to find the names that work.

List rental companies generally require a minimum of 5,000 names for a test; however, many mailers have chosen significantly higher cutoffs. “We like to test between 10,000 and 25,000 names, representing at least 20% of our overall prospecting,” says Danielle Savin, vice president, direct marketing for lingerie cataloger/retailer Frederick’s of Hollywood.

For his part, Jim Warden, circulation manager of Bowling Green, KY-based Camping World, which sells accessories for recreational vehicles, likes to test 10,000-20,000 names, while Scott Hartle, president of Columbus, OH-based western apparel cataloger Rod’s Western Palace, says his typical test size is 5,000-10,000 names.

But even as catalogers test a greater volume of names, their criteria have become more narrow.

“Testing has become much more targeted,” says Fran Golub, senior vice president, list management for Greenwich, CT-based list firm Walter Karl. Catalogers are “getting down to the nitty-gritty and cherry-picking lists” as well as requesting additional selects.

For example, a health and beauty products cataloger looking at the list of high-tech gadgets marketer Sharper Image will most likely first test select specific personal care and home item buyers rather than simply select all of the females on the file.

Targeting your tests comes at a price, of course. While catalog lists rent for an average of $85/M-$115/M, selects can range from an additional $5/M for gender to $15/M for age to up to $40/M for recency.

In addition to requesting certain selects, “brokers are asking more- detailed questions about the source of names, whether it be Internet, television, or otherwise,” Golub says, noting that consumers who purchase via different channels have different buying patterns and attributes. For example, Internet buyers tend to be quicker to purchase than direct mail buyers, while TV-generated names tend to be the most impulsive – they see it and want to order it, she says.

Beyond new selects Some catalogers aren’t content with testing new selects; they’re now testing new types of lists altogether.

“With tests, we’re always looking to come up with something new, to try a different focus – for example, magazine lists,” says Glenn Gibson, circulation manager of Allston, MA-based Bits & Pieces, a puzzles and games cataloger. “Right now, lists from industries off the beaten path seem to be working for us,” he says, though he won’t provide specifics.

Rod’s Western Palace takes a more cautious approach. “We try to do most of our tests on lists with businesses related to us, which have similar customers, but not necessarily the same product,” Hartle says.

And Frederick’s of Hollywood checks out its list-renting peers. “We look at similar lists that perform well for other companies,” Savin explains. “Let’s say the Spiegel catalog list is working well for me. I may find the other companies that are also using Spiegel and see what other lists they are using because those lists might also work for me.”

Modeling – using statistical information to build forecasting tools to predict the variables and customers most likely to respond to an offer – is a tried-and-true tactic for improving the performance of house files and large prospecting mailings. But some catalogers have started to use it when testing lists as well.

“Our service bureau has prepared a prospecting model,” says Camping World’s Warden, “through which it took our database and ranked the sales performance of customers within zip codes across the country. We then grab the highest-yielding 20% of zip codes, and when we rent outside lists, we take only names that correspond to these areas.”

Savin says that Frederick’s of Hollywood is also using modeling, to help it test new segments of tried-and-true lists: “We are going deeper within our best-performing lists. For example, rather than using three-month buyers, we are using 12-month buyers and applying modeling techniques.”

Most of the catalogers contacted test approximately two to four times a year, with holiday, followed by spring, being the prime seasons. Rod’s Western Palace, for instance, tests two to three times a year. “At holiday we might test five to eight lists. This is the time we get the highest response rate,” Hartle says.

As for what constitutes a successful test, Warden says he looks for a response rate of at least 1%. Hartle sets the stakes a bit higher: “We must get at least a 1.5%-2% response rate to move forward.”

Once a list makes the grade, catalogers begin the list rental rollout. “After a successful test, we increase our list rental size progressively,” Warden says, although “a lot of times we just go forward with the entire universe – 50,000-200,000 names.”

Hartle, who rents 50,000-100,000 names on a good run, agrees: “If the test is good, we will rent all we can.”

More privacy talk, little new action The Web has heightened consumer awareness of privacy issues, and as a result, some mailers are more careful about the print and e-mail names they rent. For the most part, however, catalogers say the Web has not changed their privacy efforts.

Tari Stanton is e-commerce manager of Henderson, NV-based chocolates cataloger Ethel M. Chocolates.

We’ve seen an increase of about 1%-2% in the number of individuals requesting that we not rent their names. This is most likely because customers are feeling the effect of privacy issues from other channels, such as the Web.

When renting online lists, we do ask more questions now about where the names are from and if the names are opt-in. On the print catalog side, we challenge the integrity of the data; if, for example, we notice business names mixed in with individual names, we may think twice about the source’s reliability.

We clearly separate our mailing names from our e-mail names, but some catalogers may combine names in the exchange. When that’s the case, catalogs that rent the names can’t see how many are Internet shoppers. We need a means of labeling these customers so that we can target them appropriately.

Michele Rick is director of customer acquisition at Charlottesville, VA-based electronics cataloger Crutchfield Corp.

In our catalog, the traditional opt-out is given prominent placement on the order form, as it always has been. Our e-mail list is absolutely not available to anyone for rental. We have always been concerned about our customers’ privacy.

When renting lists, in light of increased privacy concerns we are more sensitive and are starting to ask the broker questions about the source of names when prospecting. This change has happened fairly recently; as we have become more aware of privacy issues regarding our site, we are now more conscious when we rent other companies’ names.

In the future, I think it will come down to the industry policing itself. If companies can play by the rules, privacy issues shouldn’t become a problem. But if a few marketers don’t stick to the rules, and consumers retaliate, the offending companies could jeopardize the situation for the rest of us.

Geoff Wolf is general manager of Durango, CO-based Back in the Saddle, a catalog of horse-themed gift items.

On the printed side of direct mail, privacy was a bigger issue about five years ago; e-commerce has now taken its place as the big issue. I think printed mailings have matured to the point where most catalogers have learned to respect the names of people who don’t want to be mailed to.

On the Web, our customers have to opt in before we will send information to them. But our Internet database, which consists of a few thousand names, isn’t really big enough to share at this point, even if customers do opt in.

At some point, I would anticipate a federal privacy regulation that will be a standard for all channels of distribution, including direct mail, e-commerce, and retail.