CRM and e-commerce: After the dot-com bust

Even if you don't expect your company's Web site to contribute a substantial percentage of corporate revenue, it can still be a key strategic channel.

That was one way speakers at yesterday's CRM Summit tailored their presentations for the post-burst-Internet-bubble world. The summit was part of the eTail 2002 East conference taking place in Boston through tomorrow.

"The Web is an experimental laboratory," said Ronny Kohavi, senior director and chief evangelist of business intelligence at Blue Martini Software Inc., which, with IBM, sponsored the daylong summit. He advised attendees to test new products and marketing campaigns with small groups of online customers before rolling them out sitewide or in stores.

The Web can help companies discover what customers want in a way not possible in other retail channels. Consumer Reports, for example, looks at results of failed search requests to see what new information its readers seek. If, say, there were many failed searches for "Webcam" because the Consumer Reports Web site had no articles rating Internet video cameras, it might be a signal to assign an article on the subject.

Likewise, failed searches can tell retailers what products their customers are interested in that the company doesn't currently offer, Kohavi said. This doesn't mean retailers should run out and stock merchandise unrelated to their core missions; but companies can look to partner with other sites that do carry such products -- and get referral revenue.

A retail Web site can serve as an "early warning system for demand spikes," agreed Forrester Research Inc. analyst Jim Crawford -- if companies monitor, track and analyze the right data.

Retail sites used for research, too

If a customer fills an online shopping trolley with merchandise but doesn't go through with the purchase, has the e-commerce Web site failed? Conventional "abandoned shopping trolley" metrics would say yes, but Crawford says not necessarily.

Take, for example, a customer who put two treadmills, roughly the same price but from different manufacturers, into the trolley. Did he really plan to buy two treadmills, or was the shopper simply doing research for a brick-and-mortar purchase?

People come to retail Web sites for reasons other than buying, Crawford said, and the smart retailer won't simply focus on boosting online browse-to-buy conversion rates, but will also support the online researcher who then moves to a brick-and-mortar store to purchase.

So a store might track whether a user first put items into a shopping cart and then went to an online store locator; then, if the customer sought driving directions to a store, the printer-friendly version could include product information along with the directions, he said.

Retailers, Crawford said, need to ask, "What is a Web site good for besides selling?"

Tailor metrics to customer needs

Those seeking to analyze customer data must first understand what's important to their customers, according to Purdue University's Richard A. Feinberg.

That may seem obvious. But Feinberg has seen a number of organizations that measure business success by using metrics that aren't critical to their customers. For example, some call centers consider how much time before a customer call is answered to be their most important metric -- leading to shorter wait times -- but in fact their customers would rather wait a bit longer in exchange for having their problem solved by talking to one person (referred to as "once and done").

"What you measure determines what you get," he said.

Better information can boost revenue

Selecting what to measure can be tough online, where it's possible to collect reams of information without necessarily gleaning important trends. "The battle is not [Web site] traffic but information," said Feinberg, who is director of Purdue's Center for Customer Driven Quality.

Some experts advise collecting and analyzing as much data as possible. Feinberg, however, suggested keying in on three or four metrics considered vital to business success and concentrating on measuring and analyzing those.

One issue of almost universal importance: determining a company's best, most lucrative customers, he said. The goal is to devise different strategies for dealing with those top customers, including ways to keep them away from competitors. "[It] doesn't mean you treat your poor customers poorly," he said.

Kohavi said collecting more information can help businesses tailor strategies for different types of customers, and scrutinizing Web site activities can make a site more successful. Blue Martini worked with Harley-Davidson Inc. to analyze drop-off rates for each page of a particular e-commerce process on the site, he said. Based on the findings, Harley-Davidson revamped the process, resulting in a 120% increase in revenue from that area of the Web site.

"A lot of people sit on mountains of information," Kohavi said, "and never understand what's in [them]."

Join the newsletter!

Error: Please check your email address.

More about Blue MartiniForrester ResearchIBM Australia

Show Comments

Market Place