Predictive Analytics: The Futurists' Formula

Bulking Up CRM

At Enterasys Networks, a Siemens enterprise company that offers computer networking and wireless infrastructure and services, the use of predictive analytics has resulted in what Chief Customer and Marketing Officer Vala Afshar calls "a transformative effect" on the management of its sales cycle.

Enterasys, which uses a multitude of Salesforce.com products, such as Service and Sales Cloud and enterprise collaboration tool Chatter, wanted to find a way to calculate the likelihood of its winning a deal based on a number of indicators. Because of Salesforce.com's open APIs, Enterasys was able to develop predictive models in-house by modifying and enhancing its CRM system.

The first item on Enterasys' sales list was deal size. By combining historical analysis of deal data with "rush-in analysis of our sweet-spot deal size," as Afshar puts it, the company was able to assign weights to different deal sizes. Deal age was also a factor, and Enterasys found that if an opportunity was more than 120 days old, there was less than a 3 percent likelihood that the deal would close, which helped Enterasys prioritize which deals to go after. Then Enterasys incorporated verticals in which it traditionally performed well to close the loop.

As a result, when Enterasys sales managers have conversations with sales reps, deal size alone is not the predominant factor. "Now that we could profile these various markers, we could have a more intelligent dialogue with our field sales, and say, 'Why are you confident this deal will come in, given the fact that, historically, deals of this age have a very small likelihood of closing successfully?'" Afshar says.

Reducing Churn

Trident Marketing, the second largest reseller of DIRECTV in the United States, was tasked with improving its customer churn rate and delivering a more optimal customer to the satellite TV provider. While Trident's churn rate was between 5 and 6 percent, DIRECTV wanted it below 3 percent, or the reseller could risk losing its license and ability to sell DIRECTV.

Turning to analytics partner Fuzzy Logix, Trident Marketing uncovered mounds of customer information through the IBM PureData System for Analytics platform by identifying 30 customer variables based on nearly two terabytes of data from its telephony, CRM, and order entry systems, as well as external data from credit bureaus and clickstream data from Bing and Google.

Trident also identified which variables created a good customer, says Brandon Brown, the company's chief information officer. For example, a customer who signs up for DIRECTV to obtain the NFL Sunday ticket fan package during a marketing promotion and then cancels it when the season ends might have negative customer variables.

"We took the analytics part of [the predictive modeling] real time, and I can now inject this into the sales process," Brown notes. As a result, Trident's sales team can tailor messaging around more variables. Hypothetically, an electronics sales rep might look at a customer account and say, "You have six receivers because you have a large house and now I might be able to sell you an add-on—maybe a contract for in-house maintenance for your electronic equipment."

Within 60 days of Fuzzy Logix running predictive analysis on Trident's data warehouse, Trident noticed a 10 percent increase in sales and was able to reduce its paid search marketing costs by 30 percent. Because it had more insights into what products a particular customer might wish to buy, the company reached its ultimate goal—reducing its churn rate by 50 percent and reaching the rate target DIRECTV had set for its licensees.

Previous Page Next Page