Steps taken to build Sevita’s first enterprise data platform

Sevita is dedicated to providing adults, children, and their families innovative services and support designed to lead to growth and independence despite physical, intellectual, or behavioral challenges. With person-centered care, the company works to foster independence, improve quality of life, and promote overall well-being for the individuals they serve. As such, the data on labor, occupancy, and engagement is extremely meaningful. Here, CIO Patrick Piccininno provides a roadmap of his journey from data with no integration to meaningful dashboards, insights, and a data literate culture.

You’re building an enterprise data platform for the first time in Sevita’s history. What’s driving this investment?

When I joined in July 2022, the company had spent the prior 24 months completing more than 20 acquisitions, and the IT team was busy bringing all these new systems online. Our legacy architecture consisted of multiple standalone, on-prem data marts intended to integrate transactional data from roughly 30 electronic health record systems to deliver a reporting capability. But because of the infrastructure, employees spent hours on manual data analysis and spreadsheet jockeying. We had plenty of reporting, but very little data insight, and no real semblance of a data strategy.

This legacy situation gave us two challenges. First, while we could see some data, we couldn’t see the correlation between data points, which is where the insights lie. Second, the manual spreadsheet work resulted in significant manual data entry. We knew we had to bring the data together in an enterprise data platform.

What kinds of data did you want to correlate?

We always need insights on occupancy rates within our human services and healthcare community programs to execute our mission. As a labor-heavy business with over 43,000 employees, most of whom provide day-to-day services to the people we serve, we need better visibility to labor utilization data. Without a clear line of sight into occupancy and labor, we can’t make effective hiring decisions.

How is the new platform helping?

For the first time, we’re consolidating data to create real-time dashboards for revenue forecasting, resource optimization, and labor utilization. We’re doing KPI visualization and trend analysis, and highlighting variances over time. We also created an HR utilization dashboard to improve shift scheduling efficiency by state and by area, and  can now see how to maximize staff utilization and minimize overtime pay, rather than rely on third-party contractors to fill shifts.

For our pediatrics business, we’re using data to improve our marketing efforts to better recruit foster care providers, and to help us see where the greatest needs are by state, region, and program. This allows us to better focus our resources to match the needs of the people we serve.

But more than anything, the data platform is putting decision-making tools in the hands of our business so people can better manage their operations.

How would you categorize the change management that needed to happen to build a new enterprise data platform?

We thought about change in two ways. There’s changing the business to improve program sponsorship, the value we place on data-driven decision-making, and tool adoption. Then there’s changing IT to make sure the team is aligned, trained, and capable of managing this migration and maintaining it for years.  

In terms of business change management, the company recognized the need to solve its data problems. So while we had general business support, we then focused efforts on delivering some quick wins, showing the users the value of data access, and building-out self-service capabilities to drive more confidence in the new data platform.

Some of our executives have come from organizations with enterprise data platform investments they describe as “runaway freight trains,” costing millions of dollars to build, but only a select few using it. These leaders made sure we were being efficient with our investment. Gaining the trust of these executives was a significant hurdle to overcome.

Our IT organization didn’t have much experience working in a cloud-based environment, so it was critical to hire a leader who had been through a similar data journey. This leader helped to instill in the team a new way of thinking about data, and we hired people with new skills to work alongside our legacy team.

How did you build the first use cases?

We sought out use cases by talking to the people we knew who were frustrated with the current environment, and would partner with us on this journey. We pulled these people together, and defined use cases we could all agree were the best to demonstrate our new data capability. Once they were identified, we had to determine we had the right data. Then we migrated the data to our new data lake, and stood up the new platform.

After that, we used rapid prototyping to show our business users what we could produce, and in a quasi-agile refinement methodology, we put version after version in front of the business, substituting more relevant data for less, and ultimately developing a dashboard product they felt good about.

What baseline of data knowledge do you expect your executive peers to have?

Data literacy across the company was a challenge because, as is often the case, we were all describing our business data a little differently. Early on, we ground through creating our first data catalog, building clearer definitions of our target attributes and metrics. We now expect people to come to the data training ready to use the data catalog to help them relate those targets to their day-to-day work. If everyone speaks the same language — across all the operating groups and our leadership team — we have the foundation by which we can more effectively manage the business.

How do you know the new platform is working?

Dashboard subscriptions through the BI portal are up 400% since launch, and we’re now approaching 4,500 active dashboard subscriptions. That number might seem low, but we had fewer than 1,000 subscriptions when I joined the company two years ago. As importantly, and from a more subjective viewpoint, when I’m in team meetings or traveling throughout the company, people bring up the use of dashboards and how they’re actively engaging with data. This has helped build more enthusiasm and demand, and the company continues to resource the data platform at a higher weighting than some of our other initiatives.

What advice do you have for CIOs building their company’s first enterprise data platform?

First, make sure your new data platform capability can cover the deficiencies and gaps of the current one. You don’t want simply to replace what you have with something equivalent. Maybe you need more timely data. Maybe, as in our case, you need to consolidate data across multiple platforms or eliminate the technical debt of supporting a stand-alone, on-prem data platform. You need to know where you’re going and how you’re going to add value.

Second, we had a bunch of standalone data marts, but we’re still getting important reporting from them, so we couldn’t just stop what we were doing. We had to build the new platform and continue to deliver reporting at the same time. That’s why you have to identify a use case that you can build the new platform around.  

And you can’t take your time. You have to demonstrate the value quickly, or people will give up because it’s an expensive journey. The quicker you can identify the use case and get the new dashboards and data into the hands of the business, the better.