.png?length=1000&name=SB_Wellness(1).png "SB_Wellness(1)")
When an internal corporate health team with a promising self-help health management app was spun off into an independent venture, they had a solid plan. They focused on strong health expertise, visionary product leadership, thoughtful content designers, and informed market experts who could help identify the best path forward for steady, sustained growth. In a short time, they achieved undeniable success as evidenced by increasing subscriptions, measured improvement in subscriber health outcomes, high customer satisfaction ratings, and expanding product scope.
Like so many, the COVID-19 pandemic took this company by surprise. The instant impact was evident, their plan had not anticipated a many-fold growth in subscribers in such a short span of time. Many sought the self-help app to help them cope with new stresses heaped onto already busy lives. Subscriptions leapt from hundreds of thousands to millions seemingly overnight, yet the team rose to the challenge, scaling their self-help app and content to handle the new demand.
“They definitely knew their app and their subscribers. They stepped up when they saw the market need spike” recounts Eric Wozniak, BlueGranite VP for Health and Life Sciences. “The unexpected challenge was in maintaining the analytics insight that, to date, had enabled them to help so many people manage their health.”
Over the following few months, they saw that the analytics they had been using to ensure directional success were not keeping up. Although they had invested in a modern analytics system, its design was unable to scale with the increased subscriber data volume. Growth in report run times outpaced the increase in data. The increase in delivery time to build new reports was distressing and driving an increasing backlog of needed insight. As a result, their visibility into subscribers’ activity and resulting health outcomes was narrowing daily. Managers everywhere had begun to proliferate complex, cumbersome spreadsheets to try and keep up. Something had to be done.
“There were many competing messages how to solve their analytics challenge” continues Eric, “But it was in their DNA to make fact-based decisions. Collectively, they wanted better insight into the details of the problem before deciding what to do. That’s where BlueGranite was initially able to help.”
First, BlueGranite sought to understand both the current analytics capability as well as the anticipated need. “Our team seeks first to understand, so that we make sure our recommendations ultimately will hit the mark” explains Eric. “When it comes to modern, well-architected solutions we have a great record of success. And just as each organization is unique, so too is each recommendation.”
BlueGranite soon found that although the in-place analytics system had been implemented using a modern, cloud-based platform, the analytics processes implemented on that platform were not consistent with modern analytics best practices. Data was maintained in the central database as layers of inconsistent monthly snapshots. Data processing logic was brittle, being riddled with hard-coded values and on-the-fly data quality corrections with little overarching organization. Analysis and reporting outputs spanned multiple, overlapping, interdependent analytic and report generation processes.
“Looking at the in-place system, one might conclude that the analytics volume had naturally outgrown the initial built-for-small-scale architecture and that a new built-for-large-scale architecture was the next logical step” explains Kirtis Carpenter, BlueGranite Solution Architect, “but that would be shortsighted. A well-architected analytics solution grows with any organization over time. It scales for data volume – at a competitive cost – without the need for wholesale replacement at some point. BlueGranite understands this all too well.”
BlueGranite’s approach to modern analytics has fit-to-business phases that assure a solid path to long-term success:
“We helped them see that modern analytics could be implemented to fit their organization both today and tomorrow, whatever may come” explains Eric. “At BlueGranite we embrace growth – our approach is Think Big, Start Small. This is where we were really able to help.”
Launch
The immediate focus was to provide an alternative to the faltering in-place system. BlueGranite held workshops with product leadership and technical staff to analyze and plan the coming steps. At the same time, a series of QuickStarts were kicked off that would position the team to move forward at full speed once the overall plan had matured.
“Knowing that we would need to stage the data for whatever our planning determined, we entered the Launch phase early, with a firm foundation QuickStart” explains Chris Koester, BlueGranite Senior Consultant. “We were able to employ the BlueGranite Catalyst for Modern Analytics – our collection of technical accelerators – and after just a few days effort we had data flowing into the data lake, ready for profiling and data transformation. This early implementation with built-in Delta Lake capability (which includes a data ‘time-travel’ feature) started immediately storing the data as it appeared over time, so we could analyze it later and best understand it. We felt strongly that this would help us sort out some of the data quality challenges that were already known within the in-place system.”
At the same time, BlueGranite worked with technical staff to align roles. Then followed identification and delivery of needed training to implement, administrate, support, and secure the coming analytics solution. “This was key to assuring a successful first Use Case iteration” recalls Kirtis. “We really needed all hands available to make it work, and everyone was really eager to contribute, so that helped immensely.”
Scale
Based on the foundation established during the Launch phase, BlueGranite next led the effort to Scale the solution with an iteration on the core Use Case: subscriber engagement and resulting health outcomes. This effort started by first analyzing the self-help app data in the data lake. “The tools available in the new analytics solution made it easy to profile the data,” recounts Kirtis. “We were able to readily identify and analyze which data would need special handling as we transformed it for analysis and reporting.”
Then followed short sprints of technical implementation of focused data models that staged the information optimally for operational reports and analyses. Together, these data models and reports would restore to product leadership the insights about subscriber engagement and related health outcomes needed for operational decision making and long-term planning.
Along the way, BlueGranite helped to expand existing app development processes to grow and maintain the new analytics solution indefinitely. Among these processes were quality assurance, version control, change management, and automated deployment.
Realize
“It wasn’t all smooth sailing” recalls Kirtis. “We knew going in that the in-place system had some data quality challenges. As the implementation of the new analytics solution progressed, we learned just what that meant.”
The initial iteration on the core Use Case uncovered numerous data quality challenges. Matching between subscribers and their relationships with various entities (employers, sponsors, referrers, practitioners, insurance payors, and others) was complex and sometimes important linkages were amiss. Also, the details of the information would change over time as the self-help app that created and manage it had evolved. This showed up in the data as stratified eras where different data logic would apply. These and other complications common to quickly evolving apps were a unique, complex challenge for the team to resolve.
“As we encountered these data quality challenges, we worked closely with product leadership and the technical apps team to sort them out.” explains Kirtis. “As the project progressed, so too did our understanding and mastery of the data.”
Working together, the project team was able to identify and quantify the data quality issues in a way that the product design team could work to correct them. This was enabled not only by careful data analysis, but also by the early reports and analyses being generated by the new analytics solution. “Comparing the output of the new analytics solution with the output from the in-place system, we were seeing some significant differences” explains Kirtis. “This led to questions about what was driving the difference, and interactive exploration of the data (enabled for the first time by the new analytics solution) made that easy to investigate. We found that some of the data quality logic used by the in-place system was consistently understating the quality of health outcomes. In one case, the accuracy of the information was improved from 89.5% to 99.8%. The new analytics solution helped the self-help app itself!”
Soon the new analytics solution was in place, and the BlueGranite Recurring Services team was engaged to begin its usage lifecycle.
“Recurring Services is a continuation of the project partnership” explains James Mitterling, BlueGranite Recurring Services Lead. “We form a long-term relationship with the project stakeholders to further refine, enhance, support, and drive adoption of these transformative solutions. BlueGranite’s project and recurring services seamlessly work with our partners to deliver reliable operation and continuous improvement, growing the value realized by their investments.”
