MEASURE FOR MEASURE
By Karen J Bannan

The Internet has always been a headache for protection-of-privacy advocates. IP addresses are transparent, and clicks are recorded by the omnipotent website administrator. But for vendors, however, this represents a gold mine of market information unimaginable before the dawn of the Internet.

In fact, most companies with an on-line presence use certain metrics, or so-called web analytics, to monitor site traffic. Traditionally, a website is considered a success if it does well on the following: the number of pages viewed, the average length of time visitors spend, and the number of clicks they make. These are the most common web analytics of today.

For companies that conduct sales on the Internet, however, it has become clear that page hits, eyeballs and stickiness only present an aerial view of web traffic, but lack any correlation with the actual amount of money customers spend on a website.

To Hell in a Handbasket

Consider a new wave of web metrics - things like recency, conversion ratios, and churn - that go well beyond mere traffic numbers. These analytics zero in on site users, providing more accurate profiles of on-line shoppers and their on-line habits.

What many e-tailers are looking at these days is abandoned shopping carts. Browse-to-buy conversion rates for most on-line merchants remain microscopic - typically, 3 percent or lower. That's bad news, particularly since e-tailers are now wising up to the fact that it costs less to hold on to existing customers than it does to acquire new ones.

One vendor, NetGenesis, markets an application designed specifically to analyze shopping cart activity. The software, called CartSmarts, tracks who's going to a home page, what each visitor looks at on the site, and which shoppers actually get as far as the shopping cart. One NetGenesis customer, for example, recently reported 490,000 visits to their site in a month. Great, right? Not exactly. Fewer than half of the visitors looked at the product catalog, only 15,000 went on to the shopping cart, and a mere 2,000 actually bought something.

It's hard to make a profit off a 0.47 conversion rate. What's more, the software revealed that, of the 2,000 or so purchases that were made, most were for less-expensive products. Armed with that data, the e-tailer considered adding live customer support and other forms of on-line hand-holding to boost the sales of higher priced merchandise. In Asia, Singapore-based Creative Technologies uses NetGenesis's products to monitor its on-line shop. The company does not want to divulge whether they work or not.

One web analysis application, designed by Black Pearl, uses an artificial intelligence model. The program helps on-line sellers predict which prospective customers are likely spenders - and generates ideas to keep them around. "Based on what a customer does, clickstream data (the series of clicks made by a visitor) shows certain types of proclivities," explains Lisa Hammitt, Black Pearl's chief technology and executive officer. The AI-based program costs between US$200,000 and US$2 million - not cheap. Then again, Black Pearl sells the software mostly to financial services companies, telco providers and large manufacturers. Merrill Lynch, for one, rolled out the application this past spring. Brokers use the software to recommend courses of action, based on real-time events, to their high-net-worth customers. According to Hammitt, the software saves brokers up to six hours of research time. (Officials at Merrill Lynch would not deny or confirm that number.)

Is Recency a Word?

Scores of on-line merchants have started placing a high value on recency, the duration between visits. At the same time, stickiness, the length of stay per visit, has lost some luster as an e-metric. Why? As in the real world, merchants don't want customers hanging around all day - particularly if they're just looking. Bruce Richardson, senior vice-president of research strategy for AMR Research, points out that a high level of stickiness does not give a true picture of customer retention. "I use MyYahoo all the time, but I never buy a thing from Yahoo," he says. "They know who I am, know my age, but they are probably unable to figure out why I'm not buying, even though their site is sticky for me."

Indeed, some analysts claim it's a bad idea to put too much stock in any e-metric - even in more-refined metrics like recency and browse-to-buy ratio. "Drawing conclusions from analytics takes a lot of supposition," insists Charles King, a senior industry analyst with Zona Research, a technology research firm in California.

King says that far more people use the Internet for research than buying. "You can't simply say 42 percent of people hit this page and then abandoned their shopping cart, so there must be something wrong with the page or the process," argues King.

Still, if there's one thing CFOs do, it's measure. So it's unlikely they're going to stop using web analytics - particularly as those analytics get more accurate. Even without hard ROI numbers, one can get considerable mileage out of web analysis software. "We've been able to explain to advertisers how our revamped site is doing and why the changes we made were smart," one user claims. "The software makes it easier for us to keep advertisers and users happier."

Karen J Bannan is a contributing editor at eCFO