Homecare and Medical Staffing Company BrightStar Care Sees 10x Performance Improvement
In the homecare and medical staffing industry, BrightStar Care is a leader. With so many variables in human health, it’s a sensitive market to manage. Since BrightStar specializes in providing quality service at reasonable rates, collecting and understanding all their data is the key to their performance.
When the BI tool BrightStar previously used could not produce user-friendly reports (only cumbersome columnar reports), Jayesh Vaswani, Senior BI Developer, realized they needed a better way to visualize their large volumes of data. Jayesh’s goal was to gain more information on the company's scheduling and financial performance, as well as optimize their IT processes - which included accessing clean and consolidated data from both their internal processes and their franchise locations.
Sisense was implemented at every level in the company and now gives BrightStar access to the insights they need - leading to some dramatic results.
The Opportunity: Improving Efficiency and Care
With over 14 years in business and 10 selling franchises, BrightStar has 340 locations throughout the United States. Named by Forbes Magazine as one of the top ten best franchises in America for the third year in a row, BrightStar is a true leader in the homecare and medical staffing industry.
BrightStar wanted to measure their own IT processes, monitor their franchises performance in relation to effective scheduling and budgeting, as well as give their franchises the ability to monitor their own systems and reporting (using embedded analytics to offer customer-facing dashboards). BrightStar uses Sisense internally to better understand and manage their IT processes, and the positive impact IT optimization had on their franchises.
The Challenge: Data Volume and Ad-hoc Queries
BrightStar’s first challenge was to answer the question “how can we quickly answer business questions without rebuilding the whole pipeline every time we need to analyze additional data?” On top of that, they wanted to be able to distribute their dashboard solution globally so that their franchisees could access their own data in a few, simple clicks. Finally, they needed the ability to allow their franchisees to visualize data and provoke action as easily as possible.
BrightStar’s data was spread across their home-grown application called ABS, which was also being used by their franchisees to manage their operations. Additionally, the data was located in their accounting software, help desk software, survey systems, and Microsoft Dynamics. BrightStar needed to pull the data all together into a single data warehouse where they could analyze it in one place.
With a data warehouse of only 70GB (and tens of millions of rows), it wasn’t the size so much as the complexity of the data and the relationships between the sources that was tedious. BrightStar needed a reasonable way to visualize it. Their criteria for a BI tool was:
- Self-service BI for franchisees
- Agility of the system for ad-hoc queries
- Visually attractive presentation from graphs and dashboards
- Low cost of ownership
The Search is on
With a firm idea in mind of what they needed, Jayesh reviewed Gartner’s BI Magic Quadrant to find suggestions. He chose Sisense because it was simple to use, could be tested on their own data, and met all their criteria. The free POC and ability to grow with the product instead of an ‘all or nothing’ approach made Sisense the perfect tool for the job.
The data belongs to the business, and the business should be able to analyze it – but we wanted to maintain governance so that the business actually has the right information.
Avoiding High Traffic Issues
BrightStar wanted to get familiar with Sisense before deploying it to their users and franchisees, so they first started analyzing the usage of their own internal reports. The IT department wanted to see what reports the franchisees were using, how often and when, and relative overall usage by franchisee.
BrightStar had previously thought that Monday and Tuesday were the ‘high traffic’ days for using reports, however, it turned out Friday was the heavy day. Knowing when the traffic was highest allowed BrightStar to plan ahead for any outages or possible problems. It also allowed BrightStar to load balance batch reports better by knowing when the usage gaps would be.
Jayesh next evaluated the processing time on reports and frequency. With so much data and so many clients requesting reports, performance was critical. They’d see a report with 45,000 requests and another with 10,000. Looking into the details, they found that most of the reports had fairly similar execution times with just a few exceptions, and by highlighting the problem areas they were able to optimize the reports and speed up their execution by a factor of 10.
Finally, seeing what reports the franchisees were using allowed BrightStar to advise them more efficiently on operations and tools they might not be taking advantage of. This internal usage led to an explosion of useful ideas for development throughout Brightstar, as well as the franchisees.