An eye on Web analytics

Do you know how successful your e-business project is? Sure, top management is impressed, online sales are holding their own, and the site's architecture is sound. But do you really know which products and services your customers will buy online and whether shoppers care more about speedy delivery or low costs?

Now more than ever, solid facts and metrics are necessary when it comes to measuring the success of e-business.

Executives in many companies have sunk a lot of money and resources into various e-business projects that were driven by loosely defined promises to "improve relationships with customers" or "streamline supply-chain performance". Now, with the reality of an economic downturn, many of these projects are under increased scrutiny.

Unlike established customer channels, such as call centres and a direct sales force, Web channels have few effective metrics. The many ways of measuring this channel are known collectively as Web analytics. Among the few executives who have heard of Web analytics, most tend to dismiss it as being of interest to only the techiest of technicians.

The truth is that we will all be hearing a lot more about Web analytics in the near future. At its core, Web analytics can be thought of as real-time decision support. We've been mining customer data and sales figures within the corporate boundaries for a long time; Web analytics lets us do this dynamically.

Let's take CRM (customer relationship management), for example. Well-run companies have known for a long time who their most profitable customers are, and they have used this information to up-sell and cross-sell.

Still, prior to the Web, we could only do CRM to a certain degree. Although we could track historical purchases, we could not capture buyers' reactions to products (because it was impractical for a salesperson to follow a customer around a store).

However, with Web analytics, it is relatively simple to tell which products customers are looking at on the Web.

By offering real-time information about how products and services are selling, the Web can foreshadow customer acceptance of new products. A clothing company, for example, can prelaunch a clothing line online and use Web analytics to see how customers will respond. Online marketing campaigns can be refreshed every few hours, instead of every few months.

I know that all this talk about mining customer data sounds like a job for a database guru, but keep in mind that Web analytics is, first and foremost, an analytics exercise, not an IT project. The technology is important, but even more important is to first make sure you capture the right information.

For this reason, chief marketing officers should conduct a number of pilot programs in which "learning" is the primary objective before creating a production system. After the pilots have been completed and a production system has been set up, enterprises should consider forming a BI (business intelligence) department to centralise all kinds of analytic tasks. The necessary skills are too scarce and expensive to leave scattered throughout the enterprise.

But the skills -- or the lack of them -- are not the only issue around Web analytics. The chief marketing officer should also work with the legal department to address privacy issues. The potential for widespread customer backlash against intrusive user profiling is great. And companies should closely study trends in privacy legislation.

From a technology standpoint, the Web analytics market has been growing since 1997. The traditional leaders in the analytics space -- WebTrends, NetGenesis, and Accrue Software -- are being challenged by new players.

The direct challengers include new Web analytics product and service vendors, as well as vendors that are traditionally known for extraction, transformation, and transport of data, BI, and CRM systems.

Rather than focus exclusively on Web analytics, many of these vendors now are adding Web-usage data and common analytic functions to their product offerings.

At this time when companies are under great pressure to make every dollar count, an investment in Web analytics makes sense. The market is young and the products are immature, so avoid long-term vendor contracts.

Companies have spent decades trying to figure out how to manage decision support in the legacy world, but Web analytics is something that e-businesses can exploit immediately.

Barbara Gomolski is a research director at Gartner Group. Send her comments at BarbaraGomolski@earthlink.net.