"On the tactical side, it's straightforward," said Jim Sterne, president of Target Marketing and founding president of the Web Analytics Association. "You measure the exact response you're getting from a series of banner ads, keywords or even direct mail and newspaper ads, and determine which ones rang the bell the most."
But Sterne acknowledged the complexity of the process. One particular banner ad may elicit few click-throughs but result in higher ultimate sales. It's clear other factors are involved, but which ones?
One classic form of determining resource allocation is used in direct mail. Responses to a split-run—two waves of mailings sent out with slightly different messages or offers—are studied, leading to further refinement of a campaign.
The online version of this technique goes even further. Multivariate testing is essentially a split-run on steroids, assessing dozens of components in any single landing page, and allowing visitors to "vote" with their clicks on preferred content or offers.
"It involves continuous tweaking," said Josh Manion, CEO of Stratigent, a Web analytics company. "The good news is the returns that people are experiencing are staggering."
Just how staggering is illustrated by a Stratigent client, Designer Linens Outlet, a subsidiary of bed-and-bath products manufacturer American Pacific. Through rigorous analysis of customer behavior on Designer Linens' Web site—and subsequent Web optimization—the company was able to reduce its effective revenue share (money spent relative to money brought in) by 25%. Aiding this has been renegotiated affiliate, keyword and click-through contracts that weren't producing.
A key outcome for Designer Linens, said company marketing specialist Bev Dantz, was allocating marketing dollars appropriately.
"If a person didn't buy the first time they came to my site but went first to Google or then to an affiliate channel partner, and I pay each of these, I'd wind up paying for each partner for one sale," Dantz said. In response to Stratigent's analysis, Dantz was able to grow revenue generated per marketing dollar from about $3 to almost $6 over the course of 18 months, producing sales gains of 4%.
"Everyone has a finite budget and is trying to optimize each of his different channels, either in sales or in things like loyalty or deeper engagement," said Brian Tomz, senior director of product strategy at Coremetrics. To do this, he said, companies need to apportion credit for final sales among their marketing "touches."
For example, Coremetrics analyzes the direction sales come from and how much credit should be apportioned to each customer interaction. Customer transactions may involve searches, subsequent e-mail blasts, white paper downloads, affiliate programs, call center contact or a reviewer site that leads back to the company's product page (or channel partner) for the final sale.
"The old model of credit would attribute the sale to that final touch," Tomz said. "But the reality is, I as a company had to build the customer's trust over multiple experiences and he had to consume information over multiple sessions."
Tomz said his model gives high credit to a first mode of interaction, what he calls the "branding touch," although the system may split the credit with the last touch leading to a sale, or indeed with any others in between.
For marketers, the increasing use of technology and technological buzzwords can be dizzying. For example, National Semiconductor's VP-technical sales tools, Phil Gibson, who currently puts his emphasis on Google AdWords, is enthusiastic about so-called Web 3.0 capabilities, including dynamic client-side dashboards that would allow individuals to personalize their Web experiences. Technologies such as Adobe Flash, AIR and Flex as rich-content development tools have the potential to change the viewer experience (and, presumably, conversion) dramatically, Gibson said.
Technology aside, future success will depend on increased response-attribution metrics that rely on knowledge about customers, their habits and their responsiveness to multichannel campaigns, according to research firm Gartner Inc.
Gartner projects that by the end of 2007, Fortune 1,000 companies—using these technologies to pinpoint customer affinity for Web ingredients with the highest degree of interaction—will have increased their investment in these areas by 40%.
"The marketing department has clearly recognized that the Web is central to getting their message out," Target Marketing's Sterne said. "But when they also find out that Web analytics isn't esoteric or completely mysterious, and that it actually gives them the ability to ask smart questions, they start asking smart questions. Then, the possibilities become endless."