Businesses are getting better at analyzing customer data to determine satisfaction levels and the potential for cross-selling products. But what if companies could gain insight fast enough to stop unhappy customers before they bolted?

Online delivery service eCourier works with 2,000 clients that expect to be able to use the company's Web site to have packages picked up and delivered anywhere in central London within an hour--and none of those clients is bound by a contract. "We're only one delivery away from someone deciding to use a different delivery firm," says Jay Bregman, eCourier's co-founder and CTO.

ECourier wants to know the moment things start to change. If a steady customer begins using eCourier less, it probably means it's using a competitor more, Bregman says.

ECourier had been using Business Objects' Crystal Reports business intelligence software to run weekly reports on how much each customer booked. But by the time the company analyzed the week-old data, it couldn't appease a customer who, a week or two earlier, was upset because packages were late getting delivered. "By the time you've called the customer, they've already been with another company for a week," Bregman says. "This is a very competitive market, and if a competitor gets back a client they previously lost to us, they're really going to coddle them. That makes it very hard for us to get that account back."

INSTANT DATA

So eCourier started using software from a company called SeeWhy Software to try and generate customer data more quickly. What's unique about SeeWhy, says Bregman, is its ability to report what's happening with customers instantly.

When a new booking enters eCourier's database, the information is duplicated and saved into a repository within SeeWhy. The software then interprets the data by comparing it with previous information and trends, and if it notices an anomaly, it takes action. If a customer typically places an eCourier order every Thursday morning between 9:30 and 10 and there's been no contact during that time, eCourier's CRM team will receive an alert shortly after 10 that includes the client's history and the number of bookings it typically places in a day.

ECourier has been using SeeWhy for just a few months, but the anecdotal evidence is encouraging. "We know of 10 cases where there were problems, for whatever reason, and we got in there early," Bregman says. "As a result, we were able to save the client."

If an account manager calls the customer and learns the anomaly didn't result from dissatisfaction, the customer still appreciates the gesture, Bregman says: "It gives them the feeling that their account is being watched, and that we offer proactive customer service."

CLOSING THE BARN DOOR

There are other approaches to judging customer dissatisfaction.

Cablecom, a Swiss telecom company, used SPSS's statistical software to mine customer data, primarily from trouble tickets--such as the average duration of a ticket, or how many tickets had been opened for a customer over a specific time period--to build a model that could flag when a customer was at a high risk of leaving. But the model proved to be only about 70% accurate, says Federico Cesconi, director of customer insight and retention.

So Cesconi used SPSS's Dimensions survey research software to create an online customer survey, and from that he was able to determine that customer dissatisfaction usually begins around the ninth month of service, with the bulk of the customer losses occurring between months 12 and 14. Cesconi then created another survey that he now offers to customers in the seventh month of service, and which includes an area where they can type in specific complaints and problems. Cablecom calls customers within 24 hours of completing the survey, Cesconi says. The two approaches together, Cesconi says, provide the best view of customers ready to bolt, and the best chance at retaining them.

When it comes to customer service, no company is perfect. The right combination of BI software tools and customer-retention business processes, however, can make the difference between another chance and a lost cause.

Photo by Getty Images