|
By Tom Groenfeldt
What’s your Web site contributing to your business?
Most companies don’t know, even though they often have all the tools they need to measure what the Web is contributing to their business, says Eric Peterson, CEO of consultancy Web Analytics Demystified.
This is a great opportunity for CIOs to make a difference by bringing the resources together and providing valuable information to the company.
Some companies definitely get it -- Wal-Mart, Harrah’s, Nokia, CapitalOne, Expedia, eBay and Amazon, to name a few. And most of them won’t talk about their Web site’s analytical programs because they provide a competitive advantage. But Peterson did hear an Amazon exec talk several years ago at an industry conference. The company runs hundreds of tests on its Web site, changing colors, placements and texts to see what works best, and then makes changes constantly.
Businesses are accustomed to measuring their employees, customer satisfaction, market penetration and even the effectiveness of advertising campaigns. The Internet now makes available huge amounts of data with details that were previously impossible to measure. In a physical store, a sales clerk might watch a customer search through items and then make a selection, ask for help, or walk away.
On an e-commerce Web site, a company can collect data on which items a customer scrolled over, what they put in a shopping cart, what they removed, when they stumbled and quit the site, if they looked at one television and bought another brand, and how they responded to special offers. A single day’s Web traffic can generate more than 50 terabytes of data, says James Dietz, who manages Teradata’s professional services marketing programs.
“Soon companies that historically relied on small samples, third-party data and gut feel will have access to huge data stores of information about their customers,” says Peterson, who predicts this will power new businesses , lead to new business models and create new relationships between businesses and their customers.
With the powerful databases now available at reasonable cost, firms can maintain stores of granular data rather than rely on aggregated data or ad hoc queries. Then, using tools like Adobe’s Omniture, SAS, Google, Interwoven, or SiteSpect, they can conduct sophisticated tests on the data. Using Teradata, Oracle or MySQL or SQL Server, companies can use the data from their Web sites to improve their returns.
“The technology is by and large in place,” says Peterson. “What is lacking is the methodology and process to bring these systems together and the human capital to know which questions are reasonable to ask and then go about getting those answers.”
First-generation use of data was aimed at improving digital media, second-generation tools offered the ability to drill down and measure the results of changes made. Peterson thinks the third generation will use statistical models, forecasting, predictive analytics and decision optimization.
This is not simply business intelligence or customer intelligence, although it is related. Peterson describes a financial services firm which created visitor-by-visitor views across its retail, phone and online channels and developed extremely valuable information for its customer acquisition program.
With the information, the analytics group is presenting management with future opportunities, not past results, he adds.
Facing Resistance
Companies already have many of the resources in place to accomplish this, adds Peterson, but the data and the analysts are often in separate organizations.
Resistance typically comes from the people who have owned the silos of data for the last 25 years, he says. “They have always owned the customer record and they do research on the customer once a year. They drag their feet, but they have to incorporate Web data.”
Market research has to adjust from annual reviews to daily data collections that can be incorporated into an evolving plan. Web staff often find this is outside their comfort zone; they have to learn to deal with databases and customer segmentation if they are to add value to the process.
“The business people don’t drag their feet; they would love to have these insights, but they want them in a way that makes sense, on PowerPoints rather than on a spreadsheet,” explains Peterson. While a business will often have to make some marketing plans, such as TV advertising or print catalogs on an annual basis, with Web analytics they can become smarter about how to spend for online advertising and other marketing tools that don’t require long-term commitments.
“Here is where Amazon has a great advantage because they never got into ultra-long-term planning,” he says. “Companies with no long-term legacy are tuning their sites as often as hourly.”
Companies should test their programs, use control groups, and track the results, he adds.
“When you get good at testing, you don’t do extensive analysis ahead of time,” says Peterson. “You try different things and let the statistics tell you where the champion is.” Direct marketing companies have grown up knowing about control groups while online retail operations have often been pretty casual.
“It is common to find companies that will go and change their Web site and not bother to roll out the changes slowly and check the impact on visitor behavior,” he says. “Sometimes that blows up in your face.”
Large companies usually have the quants they need, but they are probably in finance rather than in the Web analytics group. At a Blue Cross insurer, Peterson explained what they would need in terms of data, storage and analysts on the first day; on the second day the company brought in a whole team of analysts who were employed the company; no one had thought to assign any of them to Web analytics.
“The CIO is in the exact right spot to go across the organization and find people who will step up to the plate,” he says.
Only registered users can write comments.
Please login or register. |
