Let’s say you had a business and every three months it was guaranteed to lose 100 customers.
You operated this way because you couldn’t figure out how to prevent this from happening.
And then one day someone showed you how to predict which of your customers might leave in the next three months – so that you can keep them.
The happy ending in the above scenario is now happening at Texas Children’s Health Plan. The solution is a cloud-based software platform called Predictive Analytics.
Three Health Plan departments – Finance, Business Analysis and Member Engagement – are working together to meet members’ needs through this new capability.
How does it work?
Di Miao, senior decision support analyst, explains that hundreds of data points are fed to a machine learning model to determine the likelihood of a member leaving the plan. These might include location of residence, number of missed appointments, number of providers, etc.
All of those variables are entered into a system that produces a likelihood of which members may leave. That list of members is then given to the Member Engagement team.
“When we know a member has a high likelihood of leaving, we work hard to prevent that by providing them VIP status at member events or taking extra steps to meet their needs,” said Alejandra Lima, marketing event planner. “It isn’t a perfect science but at the very least we have an idea of who might not be as satisfied with our services as we would want them to be.”
Time will tell
As is the case with most new efforts, time will tell how well the project works. However, the potential is very promising.
“I’m so excited about this project,” said Miao. “It really does have so much potential to help us care for members better. Often times when you work in positions like ours and you build the model or pull the data, you never fully understand the impact.”
But in this particular case the impact is clearer.
Miao worked alongside Kyle Stringer, senior decision support analyst, and Sadhana Sharma Luetel, data architect, to create the model.
“The old way of doing things is that you have a theory or an idea and then you use data to prove it. But when there is more data, more computing power, and more mature machine learning techniques you can take that data and then use it to discover interesting patterns or new ideas,” Miao said. “That is a better way of working and a better way of taking care of our families.”
The Next Step
The team agrees that adding more manpower to the process is the next step. “It takes time and attention to cull through the names each month and choose who to focus on. Then – maybe most importantly – it takes time to keep up with those families, follow them and determine if they actually left or not.
Adding at least a part-time staff member to the efforts is the next steps to seeing the potential of the project all the way through.