Database marketing lets you analyze sales trends and profile prospects to target top customers and give them what they want

In the summer of 1994, Mike and Brendan Moylan, co-owners of Sports Endeavors Inc., in Hillsborough, N.C., saw the other shoe drop -- and another, and another, and another. One of the buyers for their catalog company -- they sell soccer and lacrosse equipment -- had purchased 30,000 pairs of soccer shoes, expecting them to sell within two months. But by September, about 15,000 boxes of the cleated footgear were still languishing on the shelves. It wasn't until late fall that the brothers were able to unload the entire order. "That was our summer of enlightenment," says Mike. "When we were smaller, it was easy to know what customers wanted. And it didn't matter if we made a bad purchase. But by 1994, if we missed, we missed big."

That footless summer spurred the Moylans on to a remarkably productive year. With the help of a consultant, they installed a database-marketing system that could systematically and precisely analyze sales trends to help them make smarter purchasing decisions. The system combines a database full of information about customers' buying habits with analytical software that, among other things, gives buyers answers to key marketing questions: which products and colors sold best, which vendors were most profitable, which time of year was best for selling particular items, like sneakers. Armed with that knowledge, the buyers can revamp their sales approach. "With this system," says Brendan, the force behind the technology blitz, "we know we can grow."

It's a path big guys like American Express trod a long time ago: amassing large quantities of customer data in computerized form and then massaging the information to pinpoint their best customers and target them in a more personal way. In fact, through their highly sophisticated use of database marketing, big companies have come close to eliminating their small competitors' main advantage: a close, informed, flexible relationship with customers.

Yet despite the ominous signs, until recently most small companies did not use database marketing. After all, who could afford to buy and build even a simple system? And who had the time? Indeed, five years ago, when Bob McKim, a database-marketing expert, launched his Los Angeles consulting firm MS Database Marketing, he quickly had to change his focus from the Davids to the Goliaths. "With small firms, it was just too much missionary work," he says.

Now, however, small companies are finally starting to get religion, thanks to the falling costs of hardware and software and to more user friendly applications. "The combination of the power of the average PC and the sophistication of new database-marketing systems means that small businesses can have their turn," says Dennis Jorgensen, chief operating officer of the American Marketing Association, a Chicago-based trade organization. Still, especially for small players, building a system is a complicated process, often requiring outside help and a couple of years to produce a real payoff. But to hear the converts tell it, the outlay of time, energy, and money is well worth it down the road.

* * *

Know Thy Customer
The Moylans are a case in point. From the beginning -- 1984, when Mike set up shop at the tender age of 18 -- Sports Endeavors has worked hard to keep its finger on the customer's pulse. That was easy when the business was young, but as clients -- and revenues -- grew, it became harder and harder to keep track of who wanted to buy what when. That was one reason Mike brought Brendan into the business: while Mike had big ideas, Brendan had a knack for details and technology. It was Brendan who set as a priority building increasingly sophisticated database-marketing systems to follow and analyze sales and trends.

He had his work cut out for him. In 1990 the then $3-million Sports Endeavors had just one 286 PC, which it used in a distinctly low tech way: Six telephone staffers took orders manually. A keypuncher typed the orders into the PC and then printed them out and delivered them to the shipping department. "We couldn't continue like that," says Brendan.

When a colleague told him about a $60,000 turnkey mail-order management system from Nashbar/Associates called Quick/Order Processor, Brendan decided to go for it. After the software was installed on a new NCR minicomputer, everything from customer purchases to inventory status and vendors' names went into the database. But the system's analytical abilities were almost nil. "We couldn't do anything with the information," says Brendan. "We couldn't even find out what products customers were buying most."

Then came the summer of enlightenment and a startling realization: Quick/Order Processor was actually hindering the company's growth. "We needed a system that could drill down to the most minute level of data," says Mike. "To be useful it had to forecast the direction growth would come from."

This go-round, the Moylans went about the search scientifically, hiring an operations expert to find the perfect technology provider. They ended up choosing a $100,000 two-part system: a uniVerse database (from VMark Software, 508-366-3888), which they installed on the NCR, and an application called the Zircon Catalog Management System (from Zircon, 617-246-5000) to handle daily operations like shipping information and purchases. At the same time, they formed their first official MIS department to help run the thing.

But with all the fancy equipment, crucial players -- specifically purchasing and marketing staffers -- were still in the dark. That's because Zircon could manipulate the information but not analyze it. MIS had to be called in for that. If, say, buyers wanted to know how well shorts were selling, they would have to put in a request to MIS and then wait two or three days for the department to sift through the data and come up with a response. If they wanted to narrow the query down to how yellow shorts were doing, they'd have to go through the same process again. There was no efficient way for people on the front lines to tap into information that could be used for marketing and sales plans.

Enter ProBit, a data-warehousing program from System Builder Technologies (now Unidata, 800-234-5728; price starts at $25,000 plus implementation). Two years after they bought the uniVerse, the Moylans switched over to a $150,000 Data General Aviion minicomputer and hired a Unidata consultant to design a suitable version of ProBit -- one that would let buyers click on a vendor's name, for example, and within seconds see a range of data pertaining to it, or funnel down further to see, perhaps, how well the vendor's cleats sold during the past three months. The consultant installed ProBit on the Aviion and on eight 486 PCs. Now every Saturday MIS updates ProBit so that it has the week's purchasing information.

Expecting a slow beginning, the Moylans decided to focus first on forecasting. An example: based on an analysis of 20% growth in sales of Adidas watches in 1995, the company's four buyers were able to build a month-by-month sales plan for 1996.

They were also able to confirm some of their suspicions. Take suede sneakers in 1995. Buyers had noticed a decrease in demand, in favor of more trendy running shoes. But an analysis using ProBit revealed that sales had dropped by half. Now the catalog shows only a few suede-shoe styles, instead of the page of options it had featured before. Perhaps best of all, Sports Endeavors has reduced total inventory by three-quarters and sales have jumped to $25 million. "We're now able to anticipate what customers want and give it to them," says Brendan.

* * *

Pinpointing Prospects
Some businesses, such as Michael Hough's A/E/C Systems International, in Exton, Penn., use database marketing not so much to keep pace with current customers as to zero in on new ones. Hough's 16-year-old $4-million company runs computer-trade shows for the design and construction industries. Attracting attendees -- lots of them -- is its lifeblood. So Hough relies on database marketing to locate potential conference attendees and exhibitors -- primarily architects and engineers interested in computer equipment -- and to draw up promotions tailored to their interests.

Back in the early 1980s, Hough used an outside service bureau to do the computer work. The bureau maintained a list of 5,000 to 10,000 past and potential attendees, characterized by 13 attributes -- shows previously attended, for example, or primary areas of interest. Hough would draft letters aimed at customers in categories formed by combining different attributes -- for example, government engineers. Then he'd arrange for an outside mail house to send out mailings featuring sessions and materials relevant to customers' areas of expertise. "The engineers wouldn't know that the architects would be coming, too," says Hough.

Over the next 10 years, Hough made gradual changes. In 1985 he bought an IBM AT-class PC and a database-management system called dBase III, from Ashton-Tate. A couple of years later he replaced the dBase III with an early version of Fox Software's FoxPro and hired a consultant to bring the show in-house; in the process, he doubled the number of names and attributes on the list.

But then the recession of 1991 hit. Attendance at his trade shows plummeted 30%, and conference revenues dropped by half. Hough knew he had to do something fast to find customers more efficiently -- or watch his business wilt.

He turned to his database. He needed, he figured, to "slice it thinner" in order to aim his appeals more precisely. By doing that he'd also reduce the sheer number of mailings, hitting on only the prospects most likely to register.

Over the next year he refined and expanded his information base to develop more-precise promotions. Finally, in 1992 he shipped off just 25 finely tuned pitches to 500,000 potential attendees -- down a third from the year before. The cultivation of his database paid off: conference revenues rose 30%, to $300,000 (only part of Hough's total sales are derived from the conferences; the rest come from exhibitors), while promotional expenses decreased by a third.

Late last year Hough chucked the AT and shelled out $8,000 to upgrade to three 486 Windows-based machines and a faster, more powerful database-management program -- an updated version of FoxPro for Windows (now owned by Microsoft, 800-426-9400). Today his list has almost 50,000 names and 39 attributes. He uses it to send out 30 to 40 mailings, or 500,000 pieces a year. Thanks to his more systematic approach, promotional expenses continue to fall. And conference revenues continue to go up: last year they reached $600,000, and Hough expects them to top $750,000 this year. Most letters are processed by the outside mail house, which offers a discount on bulk mail of about 2¢ a piece. Hough gives the mail house a disk with names, addresses, and attributes, along with instructions about which attributes to combine and which letters to send to which combinations. "If we hadn't automated, attendance and revenues would have been dropping steadily," says Hough. "Instead, we're growing nicely and expect to continue to do so."

* * *

Sharper Distribution
Lots of small businesses don't sell directly to their customers. But that doesn't mean they can't use database marketing to make sure their distributors target the right people. Integrating sales and distribution through technology, though, is no day in the park. Just ask Agustin Huneeus. In 1991 Huneeus, president of $25-million Franciscan Estates Inc., a winery in Rutherford, Calif., set out to track and organize his distribution system. Five years later, the overhaul is still a work in progress.

An unsettling observation sparked the project: Huneeus had noticed that when he went to a fancy restaurant to dine, he rarely found one of his wines on the wine list. "Where on earth is my product being sold?" he wondered. He took to calling the phenomenon "black-hole marketing." He was producing wines that seemed to disappear into space.

Taking control of the matter, however, was tough, given the nature of wine distribution. Like any winery, Franciscan legally couldn't sell directly to customers; it had to go through distributors. To control where his product went, Huneeus had to control 50 distributors: knowing which stores, restaurants, hotels, and nightclubs they were selling to; tracking those sales; and then, as Huneeus says, "offering them guidance on how best to change their strategies."

Thus was born an ambitious plan: Huneeus would push distributors to focus on "pillars," the 1,000 top-selling stores in the country, and "windows," the 1,000 highest-profile hotels and restaurants. To track those accounts, he would hire someone to develop a computerized system that included both a database and analytical software. His salespeople would use the system to analyze at which pillars and windows which wines were selling and pass the information on to the distributors. It was an indirect database-marketing system, one that would help distributors pinpoint the best customers and sell them the most appropriate wines.

To start, Huneeus asked his distributors to produce monthly reports on which wines were sold to which accounts. That was easier said than done. Getting the distributors' MIS people to write up reports in an agreed-on format was a major headache. Most of the distributors eventually started producing reports, but the reports differed wildly. Some included monthly data; others, everything sold that year. Most listed all accounts, not just pillars and windows.

Back at headquarters, Franciscan staffers laboriously keypunched the data into Paradox, a database-management program from Borland International (408-431-1000, $284) installed on a 486 PC. Not only was the project cumbersome, but in short order it started to lose momentum. Because no one had been assigned full responsibility for it, no one followed it through.

Huneeus knew he had to do something to pull the system into shape. So he hired a new MIS manager, Alan Atkinson. Right off, Huneeus had Atkinson help customize an Excel spreadsheet program for tracking monthly and yearly sales by account, brand, and product. Atkinson distributed the application onto laptops, which he parceled off to the company's 10 regional salespeople. They in turn were assigned to undertake another laborious process -- collecting written reports from distributors, filtering through the reports for pillars- and windows-account information, and manually keying that information into the spreadsheet on their laptops. After that, they'd send an electronic copy of the spreadsheet to headquarters to be kept on file.

Through all those permutations, Huneeus continued to hone the pillars and windows concept and to search for the best computer system. (The Excel setup was always meant to be just a stopgap measure.) Finally, after a year he and his son, Agustin Francisco, who was in charge of marketing, found what they were looking for: a $25,000 (base price) proprietary system, Trade Account Management, developed by MFK Systems (707-963-9222), that was customized to work with GoldMine, an off-the-shelf contact-management program from Elan Software (800-654-3526, $200). Trade Account Management included a central database on a Micron file server; GoldMine was tied into the system and loaded on salespeople's laptops.

Here's how it works: Distributors send their reports to the database electronically. Trade Account Management breaks out the sales numbers and dates for each distributor's accounts and arranges them by region, state, zip code, and type of wine. Huneeus can then mine the system for intelligence like how his Chardonnays sold in 1992 or how his Cabernets did in the Midwest. GoldMine, on the other hand, houses the demographic information -- which retail account is buying which wines and how often -- and whether the account is a pillar or a window. By dialing into the Micron file server, salespeople can retrieve or send updated demographic information whenever they need it.

That's the concept. Reality is another matter. For everything to work, the distributors have to be conscientious about sending in reports. So far, 10 have complied, but only 7 of them supply disks in a format that's compatible with the system. In the meantime Franciscan salespeople are keying report information into their laptops, as usual.

The moral of the story? Database marketing can give you remarkable powers of perception. But at the start, patience and persistence may be the most important tools at your disposal.

* * *

Anne Field (annearf@aol.com) is a freelance writer based in Pelham, N.Y. She specializes in stories on small business and consumer affairs.


FIVE STEPS TO SUCCESSFUL DATABASE MARKETING

( Pinpoint your objectives. Are you looking to increase your customer base? To double sales of a product line? To encourage repeat business? Make sure you understand your goals before you automate. "Companies that are drawn to the technology without knowing what they want to get out of it will be very disappointed," says Andrew Deutch, of the consulting firm David Shepard Associates, in Dix Hills, N.Y.

( Choose the software that best fits your goals. If, for instance, you need a system that's tied tightly to sales-rep activity in the field, opt for a contact-management program with database capability. On the other hand, if you do a lot of telemarketing, you may need a telemarketing program.

( Determine who should be in the database. First, figure out what types of customers have the most potential. Then construct the database, using everything from salespeople's contacts to lists bought from outside suppliers.

( Develop a plan. Only after your objectives are laid out and your database is constructed is it time to devise a marketing program. Generally it will fall into one of three categories: direct marketing, rewards for repeat purchases, or relationship-building promotions for long-time customers that generate profits.

( Measure results. Generate periodic status reports in which you determine items like cost per contact and cost per sale. If you want to be really careful, don't launch your program at full tilt. Start with a small prototype; if you like what you see, expand to include the entire database.


RESOURCES

There are many organizations and books you can turn to for more information. These are some of the best:

Organizations
American Marketing Association
(800-AMA-1150, http://www.ama.org), 250 South Wacker Dr., Suite 200, Chicago, IL 60606-5819. Provides books, seminars, and publications on database marketing. Local chapters also offer their own services.


Direct Marketing Association
(212-768-7277, http://www.the-dma.org), 1120 Avenue of the Americas, New York, NY 10036-6700. Provides seminars and publications on database marketing. Also will hold two conferences this year on the subject: July 17Ñ19, in Chicago, and December 11Ñ13, in Orlando.


Books
The Complete Database Marketer,
by Arthur M. Hughes (Probus Publishing Co., Chicago, 1991). Details how to build a marketing database and use it both to find new customers and to build customer loyalty.

The New Direct Marketing: How to Implement a Profit-Driven Database Marketing Strategy, by David Shepard Associates (Dow JonesÑIrwin, Homewood, IL, 1990). A step-by-step guide to using database technology. Includes directories of data sources and database software.

Strategic Database Marketing, by Rob Jackson and Paul Wang (NTC Business Books, Lincolnwood, IL, 1994). For those who know something about database marketing. Provides insight into successful strategies.