EMC didn’t grow to be a $25 billion global technology leader without a keen focus on customer satisfaction. In fact, EMC has dedicated a program called Total Customer Experience (TCE) to drive continuous innovation in enhancing customer experience. For example, one strategy has been for our customer service organization to conduct surveys on tens of thousands of service events each month. But with only 1.2% of surveys returned, we knew we were missing some important feedback.
Enter Brad Barker, Consultant Customer Advocate for EMC’s Voice of the Customer program. Through valuable insights derived from a data lake, Brad developed the Customer Services Predictive Follow-up Program as a new way to identify and connect with potentially dissatisfied customers. To support this week’s global celebration of TCE, I had the opportunity to speak with Brad about the impact this new program is having on customer satisfaction
1. What is the Customer Services Predictive Follow-up Program?
Simply put, the Customer Services Predictive Follow-up Program predicts customer dissatisfaction. It uses our survey data, which tells us where we performed well and where we didn’t, and applies that knowledge against the attributes of each service event we handle. Then, using a predictive model built from our data lake, we can determine the likelihood of a particular service event resulting in dissatisfaction.
The customer follow-up also is important. If the model predicts customer dissatisfaction from a service event, we require the responsible manager to call the customer and attempt to resolve the issues.
2. What business drivers led to creating the Customer Services Predictive Follow-up Program?
There’s an industry standard measure of customer loyalty called the Net Promoter Score—NPS. In the last four years, EMC tripled our NPS score, from 13 to 39, which puts us in the leader category. Every VP in our company is measured against EMC’s NPS score. Since nothing can drag down that score faster than poor customer satisfaction, that was a major driver for this program.
3. What insights were discovered and how are they delivered to service managers?
We analyze about 86 different data elements, drawing on 15 years of survey results and vast metrics collected on service calls. We apply about 500 business rules in our analysis, which accurately predicts when dissatisfaction may occur during the service delivery process. For example, we know that if a time to resolution exceeds three or four days, customer satisfaction drops significantly.
We created a system that automatically triggers notification to the service manager when specified thresholds are reached. This allows the service manager proactively contact the customer to try and prevent or minimize any dissatisfaction.
4. What has the response been from service managers?
Service managers are telling us this is a valuable program. About 81% of participating service managers say their customers appreciate the follow-up call. In addition, 76% feel the program helps increase customer satisfaction. So we’re very pleased with the response.
5. What business results have you seen from this program?
There are several ways we measure success. One is increase in survey participation, and for EMC organizations participating in the program, customers returning the surveys rose 14.2%.
The second measure is change in the number of responses to the survey. We now get an increase of 12.3% responses from participating versus nonparticipating EMC organizations.
Our overall customer satisfaction, or CSAT, score is about 1% higher, but you have to consider that EMC achieves extraordinarily high CSAT levels already.
The last thing is satisfaction with EMC customer services, which is part of our loyalty program. That increased 14% since starting the Customer Services Predictive Follow-up Program, reaching an all-time high in third quarter 2015.
6. How important is having the right technology, people, and processes?
There’s no question we would not have been successful without using a data lake that merges together all the data we collect from our call management system and surveys. Our data lake centered around Pivotal Big Data Suite is critical for that.
You have to have the right people and skills. Knowledge of customer service processes is essential, along with statistical analysis skills. We partnered with an outside consultant to develop our predictive model using Hadoop and tools like the R programming language.
Process also is key. When the predictive model identifies events that may cause customer dissatisfaction, it feeds another program that alerts the service managers and generates reports that initiate corrective action. Then after the service manager calls the customer, they record their findings in the system so we can do continuous business practice improvement based on those results.
7. What would you say to other companies about improving customer satisfaction?
Programs like ours are where companies have to go. They need a way to predict how customers feel about the company. There’s a lot going on in the industry about personalized products and services. You can’t do that without understanding what makes your customers happy and what causes dissatisfaction.Tags: big data, big data analytics, business data lake, customer satisfaction, customer service, EMC, sentiment analysis, source:bdb