Segmentation categorizes individuals based on shared characteristics like demographics, behaviors, or attitudes.
This helps marketers tailor services and communications to meet the distinct needs of each segment, enhancing customer satisfaction, increasing loyalty, and driving revenue growth.
Project Goal
In collaboration with a healthcare provider specializing in workplace onsite and near-site services, we led a pilot project to develop their first-ever patient segmentation. The primary aim was to enhance service promotion among employees, focusing on increasing service uptake among currently engaged employees and acquiring new users from the pool of eligible yet inactive employees. The overall goal was to optimize health service utilization and improve overall employee health engagement.
Methodology
The segmentation was meticulously crafted by integrating multiple data sources, including the client's encounter data, employee demographics, third-party data from Experian, and external claims. The challenge was to construct a segmentation model that effectively categorized employees all employees across their lifecycle —from currently active users to those eligible, but yet to utilize the services. To create this model, R code was developed to apply advanced clustering methodologies, ensuring precise and scalable segment definitions. This coding approach allowed for the dynamic manipulation of cluster variables and sizes, making it possible to fine-tune the segmentation in response to emerging patterns and insights. Further depth was added through an employee survey distributed after segment identification. This helped refine the segments with additional insights into healthcare attitudes and usage barriers. This comprehensive method ensured that each segment was not only statistically robust, but also aligned with practical business needs and consumer behaviors.
Results
The project successfully identified seven distinct segments spanning across the client's employee lifecycle. The client was highly impressed with the segmentation’s clarity and immediate applicability. They said the segmentation presentation was easy to understand and engaging. Additionally, the data team loved the comprehensive data dictionary provided alongside the data file, which included both technical and business definitions of model attributes. This clarity facilitated internal communications and furthered understanding across the client’s teams. Encouraged by the project's success, the client extended our engagement, signing an additional scope of work to assist with the creative execution based on the segmentation insights.
Analysis Snapshot
Checkout my R Shiny Dashboard!
Working with a lean team means getting creative with the tools we use. That’s why I built this handy R dashboard, using dummy data to show you how it works. It’s designed to make our segmentation process smoother and smarter. With this dashboard, I can tweak model inputs, play with cluster sizes, and even switch up clustering methods on the fly. It's a great way to see how changes affect our model in real time, helping us nail down the best approach.
Note: This was created for internal data science use and was not intended to be client facing.