Putting the Lid on Churn

In the intensely competitive telecommunications industry, a fundamental tenet is that it costs much less to retain an existing customer than to acquire a new one. The challenge, of course, is to identify customers who are at the highest risk of “churn” before they switch to another carrier. This is something that Virginia-based XO Communications does extremely well, leveraging the power of IBM SPSS Statistics and Modeler to achieve significant improvement in customer retention and revenue protection with a small in-house staff.

Spotting potential defectors
It all started when XO Communications, already engaged in a full-scale transformation to increase business efficiency, received confirmation from a consultant that predictive analytics could be helpful in reducing churn. Trent Taylor, director of customer intelligence, jumped into action.

“At first we outsourced the analytics to an external vendor that used both SAS and IBM SPSS,” he recalls. “But we always knew we wanted to bring the function inside, and we quickly realized that IBM SPSS was the solution suite we wanted to standardize on. For us, IBM SPSS is clearly superior in terms of functionality, scalability, flexible licensing agreements, and ease of use.” Following the industry-standard CRISP-DM process, Taylor and his two-person team started with Statistics and subsequently added Modeler to their analytics tool kit.

Sub-$1,000 business customers represented the primary target for predictive analytics. XO Communications serves thousands of these customers, and while their margins cannot justify the attention of a dedicated client services manager, they nonetheless represent significant business and revenue for the company. Using IBM SPSS predictive analytics, XO and the SPSS team worked together to create an internal data mart, evaluated more than 500 potential variables for predicting voluntary customer defection within 90 days, and built a multiple regression model on the 25 most relevant variables. Monthly churn risk scores were assigned to all customers, with Client Services executing proactive communications with the top 10 percent.

“By enabling our client services managers to prioritize their proactive outbound calls – basically, a ‘health check’ on the customer – we can cover more risk with our existing Client Services team,” says Taylor. “It’s been a very successful business model for us and has helped us organize our resources better.” Thanks to IBM SPSS, each manager can effectively monitor churn risk on up to 400 accounts; the company would otherwise need twice as many client services managers to achieve the same level of churn intervention through frequent contact and relationship building.

Goal exceeded, revenue protected
The churn reduction project has paid off handsomely. According to Taylor, “We have actually exceeded our initial goal.”Comparing the retention rate of at-risk customers who were proactively contacted by Client Services to a control group, the team estimates that it’s achieving a 60 percent improvement in revenue retention rates which translates into millions of dollars in annualized revenue protection. Says Taylor: “The key difference was using IBM SPSS to prioritize the Client Services workload.”

Not surprisingly, this success has sparked interest across the company. “We have certainly proved the worth of predictive analytics over the last couple of years, and we’ve really gotten a lot of attention at the senior level as to what we can do with predictive analytics,” says Cris Payne, senior manager of Customer Intelligence.

“We have a project list as long as our arm for the remainder of this year and into next year – and with the quick workflow that you can build in IBM SPSS, we know we can handle the load.” Because churn reduction models can now be produced so quickly and efficiently, XO Communications expects to start shortening the update cycle to quarterly – or even monthly – to further enhance the interconnection of timely critical customer data across the system.

Training for best practices
The complete IBM SPSS solution included comprehensive training. Says Payne: “Most people have some sort of statistics in their undergraduate or graduate work, but that’s mostly focused on descriptive statistics, linear regression, and maybe a little forecasting. Predictive analytics is still an emerging field, and people are just starting to grasp the power of it. The IBM SPSS training is extremely valuable, because you don’t just learn the programs, you also learn a lot of best practices. The instructors are absolutely top-notch, and you can use them as a sounding board for exploring or refining your current methods. Along the way, they share a lot of tricks about fine-tuning the models.”

Within 2–3 months of IBM SPSS Modeler training, the Customer Intelligence team was already building more effective models for churn analysis then any from their previous third-party consultant. “The training helped us ramp up quickly and get the results we needed,” continues Payne. “Before taking the project in-house, we had 10 or 12 people working on it. Now it’s just myself and two other people part time, and we quickly got to the point where we were outperforming the models produced by our external vendor. In light of the revenue protection figures that we’ve calculated, we easily paid for the training in the first two months.”

Payne looks forward to the upcoming IBM SPSS classes. “We’re going to take advanced techniques in regression, clustering and association models, syntax and macros – and that’s just the beginning,” he says. “If you don’t keep up with the training in this field, you quickly fall behind.”

A successful project
The use of Statistics and Modeler within XO Communications is likely to grow. Some ideas on the table: embedding predictive analytics in operational systems, in order to suggest effective options for at-risk customers, identifying customers that would be especially responsive to targeted marketing campaigns, and modeling customer lifecycles. The team has already been contacted by other departments (including Collections and Small Business Services) for help with related initiatives.

Concludes Payne: “It has been a good journey with IBM SPSS for the last couple years. It’s gotten us to a place that we wanted to be, and we’ve exceeded our expectations with regard to stemming churn. By turning customer data into real intelligence we can use, Statistics and Modeler have definitely proven their value here at XO Communications.”

About IBM Business Analytics
IBM Business Analytics software delivers complete, consistent and accurate information that decision-makers trust to improve business performance. A comprehensive portfolio of business intelligence, predictive analytics, financial performance and strategy management, and analytic applications provides clear, immediate and actionable insights into current performance and the ability to predict future outcomes. Combined with rich industry solutions, proven practices and professional services, organizations of every size can drive the highest productivity, confidently automate decisions and deliver better results.

As part of this portfolio, IBM SPSS Predictive Analytics software helps organizations predict future events and proactively act upon that insight to drive better business outcomes. Commercial, government and academic customers worldwide rely on IBM SPSS technology as a competitive advantage in attracting, retaining and growing customers, while reducing fraud and mitigating risk. By incorporating IBM SPSS software into their daily operations, organizations become predictive enterprises – able to direct and automate decisions to meet business goals and achieve measurable competitive advantage. For further information or to reach a representative visit www.ibm.com/spss.