InformationWeek: The Business Value of Technology

For companies with multiple sales channels, combining and analyzing data from retail stores, catalog operations, and Internet sales can provide marketers with a more complete view of a company's customer base than analyzing online sales data alone. It's the only way, for example, to determine if sales through the Web are cannibalizing other channels or if they represent incremental sales growth.

But as Specialized Bicycle and CVS.com have discovered, such data may reside in disparate systems or be stored in incompatible formats. Analyst Eckerson describes pulling together such information as "an order of magnitude more complex" than simply collecting and analyzing clickstream data.

3Com Corp. plans just such an effort to understand how sales of its networking equipment, Palm organizers, and other products through its Web site are related to sales through other channels, says Ari Bose, VP of E-business and emerging technology. 3Com is already using Informatica Corp.'s PowerCenter software to collect clickstream data from its Web site's log files and load it into an Oracle database. Using query and reporting tools from Business Objects SA, marketing managers study the data to understand the usage patterns of Web-site visitors. Those findings are used to manage the workload of its Netscape application server, as well as to enhance the site's content. 3Com, for example, has been adding to the site's product configuration capabilities in response to customer usage patterns.

3Com is also using Informatica's PowerMart to pull encrypted sales transaction data from an Open Market Inc. Transact commerce engine, convert it, and load it into a Hyperion Essbase database. When clickstream data is added to Essbase, 3Com managers can correlate the behavior of Web-site visitors with information about those who actually make a purchase. 3Com marketers can determine, for example, what Web-site content is most likely to entice visitors to buy and which pages lead to abandoned shopping carts--all of which, Bose says, "drives our marketing programs."

The next step, now in the planning stages, will combine Internet sales data with data from other channels, including the company's direct sales force and retail channels. 3Com already has a data warehouse--based on an Oracle8 database running on Sun Enterprise 10000 servers--stocked with data from its traditional channels. "We need to marry that with the Web traffic data," Bose says. "That's what we're working on right now, how we can collapse all this information into the single data warehouse." 3Com plans to use PowerMart for that project as well. Bose says the development work will cost less than $100,000, not including software licenses.

Once the work is complete sometime in June, 3Com marketers will be able to understand what products are sold through what channels, and tell whether online shoppers are new customers or already buy 3Com products through other channels.

Another company with a "multi-channel" E-intelligence challenge is Patagonia Inc., a Ventura, Calif., manufacturer of clothing and equipment for outdoor activities. Patagonia has been selling its Gore-Tex jackets, Capilene underwear, and other products online since 1997. In addition to its Web sales, Patagonia also sells its products through its catalog, its own brick-and-mortar stores, and through other retail chains.

Patagonia has one database that lists its online and catalog customers, including names, addresses, phone numbers, and E-mail addresses. But data on the specific products they purchase and how much they spend resides in another database. That means Patagonia marketers can't match information about the interests of Web-site visitors--based on what content they view--with data about what they buy. They also have no way of determining if online shoppers have simply switched channels or if they represent new business.

Patagonia now receives clickstream reports from its UUnet Web-host service, with information such as where visitors to Patagonia's Web site come from and how long they stay, Web manager Craig Wilson says. But starting this month, the retailer will use Personify's software to link information from its customer and sales databases for online and catalog sales. Data from Patagonia's retail stores will be added to the mix later this year.

The convergence will let marketers match customers' product tastes with the content they view, says Morlee Griswold, Patagonia's direct marketing director. Although Personify can identify individuals and their content and product preferences, Griswold says, Patagonia's focus will be identifying trends, content, and product preferences among groups of people. The information will be used to design new content for the Web site. The E-intelligence will also help Patagonia marketers design more-targeted product-marketing E-mail campaigns for registered visitors.

Marketers will also be able to judge the effectiveness of cross-promotional efforts with partners such as American Express and Yahoo. By examining data about Web-site customers who also buy Patagonia products through other channels, the retailer will be able to gauge the degree of channel cannibalization--critical for inventory management, and for making decisions such as what products to market through which channels and how and where to spend advertising dollars, Griswold says.

While the technological hurdles to E-intelligence may be great, so are the potential rewards. New tools and technologies are shrinking the hurdles, as the competitive nature of E-commerce is only increasing the value of E-intelligence. Not since the days when neighborhood shopkeepers knew buyers' habits from experience have companies had the opportunity to know their customers as well as they can today. Says Patagonia's Griswold, "It's the kind of stuff that marketers have always dreamed of."

return to page 1, 2

Illustration by John S. Dykes
Photo of Bose by Gary Parker
Photo of Patagonia people by Edward Carreon