Neurotic Over Subscriber Numbers

Wireless carriers have a ways to go before they wrestle customer churn down to acceptable levels. Cognitive engine technology can do some of the thinking for them, but, ultimately, it's a strong customer relationship that's important.

Wireless subscribers are fickle. On average, carriers lose one out of every three customers to a competitor every year. Sure, service providers' ledgers inadvertently benefit from all this churn — at least in terms of the number of new sign-ups they get to report. But such numbers actually represent a false positive as carriers continue to hemorrhage customers at alarming rates.

With the constriction of the financial markets, Wall Streeters have gotten wise to churn's effect on subscriber numbers and the bottom line. Now they are forcing carriers to get smarter about how to hang on to customers.

Lightbridge's Pam Reeve has just the technology to help do that. Cognitive engine technology, or artificial neural networks (ANNs), gives machines the IQ to master large volumes of data. Through complex algorithms, they can process and “think” like humans (actually, they can think better than humans because they never get tired). Used in an array of applications, ranging from allowing diabetics to noninvasively monitor their glucose levels to detecting the ripeness of fruit, ANNs can deconstruct the churn phenomenon and construct a strategy for customer retention.

“The sophisticated mathematical models allow for analysis and discrimination when looking at large data sets and allows for discovery of patterns in that data that would be difficult for any human to discern,” Reeve said.

This cognitive model is just the latest addition in predictive technology that Reeve, president and CEO of Lightbridge, has amassed in her company's arsenal to fight carrier churn. In an agreement with Melbourne, Fla.-based Computer Science Innovations, Lightbridge can apply ANNs to the mountains of subscriber data spewing from carriers' networks. Plug in data, and out comes answers.

After finishing her studies at Harvard and the University of Georgia, Reeve planned to spend only a couple of years at Boston Consulting Group to get a postgraduate taste of working on big-company strategy issues. The strategic aspect was compelling enough to make Reeve stay in the field for 11 years, but readiness and opportunity finally intersected for her in 1989, the year Lightbridge was founded.

In the emerging mobile communications field, Reeve discovered that activating cellular service was akin to applying for a home mortgage. She knew it didn't have to be that complex, so she decided it was time to scratch her entrepreneurial itch. Reeve figured she could apply her previous experience to building a portfolio of products and services that would help new cellular carriers sign up customers quickly yet still collect enough data to fend off fraudsters.

Her calculations proved accurate. Under Reeve's leadership, Lightbridge grew from $1 million in 1991 to $177 million by 2001.

“We started with the very basics of how you take a potential subscriber and evaluate them from a credit perspective,” Reeve said. “If this was going to be like other industries, it would be all about getting customers and having relationships with them.”

That approach, along with predictive modeling, has evolved from being a blunt instrument to a more sophisticated differentiator with ANNs. This cognitive engine technology initially will be incorporated during the second quarter into Lightbridge's other offerings such as its Fraud-Buster and retention products. Eventually, Reeve expects the technology will spark new ideas to help carriers use customer data to shape new customized services.

“When you have an individual behavior that you want to understand, you have a body of modeling that has been done so you can make a determination about an individual behavior,” Reeve said.

This cognitive capability also can help carriers target service programs, upsell or communicate with the subscriber base.

ANNs help carriers in two distinct ways. The more carriers know about subscribers — even at the point of activation — the more it helps carriers determine the structure of the offer, what pricing point makes sense and what package of services to put together. Wireless data and other new offerings only will complicate the offer.

In what Reeve calls “customer for life cycle” management, carriers can use ongoing analysis to understand what is happening in the subscriber base.

“If there are changes going on, make sure there are outreaches, programs — whatever makes sense for a particular type of subscriber, as opposed to something that's very broadcast,” she said. “This allows carriers to outreach with more predictability and more power than before.”

Outreach with ANNs is important, but carriers also need to change their organization, people and point of view, Reeve said. Customer acquisition has moved from a transaction to a relationship.

“It's not something you plop in and do as a one-time project. It requires a different mindset in how you think about the subscriber. The technology allows carriers to deal with customers with a greater sense of nuance.”

Recalling her time at BCG, Reeve said this shift is typical of a maturing industry. After the growth of a pent-up market, “you have to move to more of a relationship approach. That is the only way to get revenue growth and profitability because you can't rely on new adds anymore.”

While maturity and financial pressures force carriers to move away from transactions to relationships, it won't be long before carriers beg for even more information about their subscribers. Being able to deploy ephemeral data services cost-effectively and managing them once they are deployed will be pressure points for carriers.

In “Machine as Mind,” noted statistician and ANN specialist Herbert A. Simon said that the human mind does not reach its goals mysteriously or miraculously.

“Even its sudden insights and ‘ahas’ are explainable in terms of recognition process, well-informed research, knowledge-prepared experiences of surprise and changes in representation motivated by shifts in attention.”

Once these are incorporated in the ANN theory, he said, the explainable is explained.

“I disagree that we'll reach a point where there will be no mystery about human behavior. I also don't think we'll ever have a world where you won't have to touch anyone or talk to anyone,” Reeve said.

That just wouldn't be very smart.