Is My BI Environment Suited to Enable Data Discovery?
Having data isn't enough. You need to make it easy for decision makers to access, manipulate, and interpret the data.
By David Cole, Chief Technology Officer, Claraview
Business intelligence tools in healthcare have traditionally not been helpful for analysts trying to determine why things happen. Data can come from multiple sources with differing levels of cleanliness and granularity. More recently, analytic software has started to place a greater focus on data discovery. Analytics and visualization tools let you dig deeper, look at likely chains of events leading to outcomes, and predict outcomes of chains of events that are still unfolding.
In healthcare, this means deeper measures of the quality of patient outcomes, such as reducing 30-day readmissions, reducing supply costs by ensuring the appropriate suppliers are charging the correct contract prices, and promoting efficient patient flow through facilities (i.e., by optimizing operating room turnover and utilization), among other measures.
Data discovery tools make it easy to navigate small amounts of data in visually appealing ways. They allow you to identify trends, small and large, as well as outliers. By visualizing data, analysts can make those "Aha!" discoveries that lead to real changes in how a business is run.
Preventing Guru Bottlenecks
There are times we find ourselves in holding patterns, waiting for IT or analytics gurus to sort through massive amounts of data and deliver reports, many of which will be out of date by the time they're generated. No one's blaming the gurus; they're doing great work with the tools they have, but traditional methods aren't sufficiently responsive to stay ahead of trends, much less predict them.
This is where data discovery tools come in. Because they can integrate multiple sources of data, analyze it quickly, and present the data interactively for further drill down, data discovery tools help users investigate results rather than just read reports.
For example, suppose you're looking at a clinical supply chain and your stent costs are higher than you predicted. Something's off. Traditional BI tools will tell you this, and that's often the extent of what they'll tell you. BI is good at that, but when it comes to asking why, BI has historically failed to address causes. Data discovery tools show you the data visually (no more mind-numbing spreadsheets) in ways that let you play with it, and by you, I mean you. Data discovery interfaces are user friendly; they take the analysis off IT's plate and let virtually anyone look for patterns and trends.
Assessing Your Environment
The tremendous promise of data discovery means that many organizations seek to leverage it without assessing their own readiness. To make most effective use of these tools, organizations should investigate their business environments and answer the following questions:
What are your problems or questions? Which problems are you seeking to solve? Which questions are you setting out to answer? It's important to identify these issues so you can appraise whether and how analytics are likely to provide effective solutions.
What are your key data sources? Identify key data sources you expect a data discovery platform to leverage to answer the questions you've identified. You may wish to consult with your data gurus (if you have them) to understand how effectively you will be able to analyze this data.
What are the best tools for your needs? Once you know what kinds of data you need to manipulate, you will have the necessary information to investigate data discovery tools. Again, consult with experts where possible to find the platform best suited to the types of data you're analyzing and the types of solutions you're seeking.
Do you have a data discovery strategy? Simply adopting a tool isn't enough. Formulate a data discovery strategy encompassing your key data sources, as well as an outline for how the tools will be applied -- and by whom -- within your organization. How will actionable intelligence be evaluated and used? Ensure that you're not investing in a tool you are not prepared to implement.
Are you prepared to train your team? One of the key advantages of data discovery is its accessibility to all manner of experts within your organization. Though these tools are highly intuitive, they will nonetheless require training to analyze and interpret data as effectively as possible.
Once you have answers to the first three questions -- and can answer "yes" to the last two -- you can move forward with the confidence that your environment is suited to data discovery.
From Gatekeepers to Enablers
With data discovery, IT personnel shift from gatekeepers to enablers. They're freed from some degree of reporting and data analysis, but it remains crucial for IT professionals to educate business analysts in the use of these tools as well as how to find the data resources they need within the business analytics environment.
There's still a big place for BI specialists, IT personnel, and analysts in the data discovery world. It's up to them to point everyone at the data, show them the level of interactivity of the tools, and teach everyone how to be as self-sufficient as possible. The better your IT department equips and trains the rest of your company, the greater the potential insight from those employees.
This is vitally important. Part of the beauty of the gatekeeper-to-enabler shift is that potentially all of your staff (not just your analysts) are able to see your data and manipulate it. Anyone with a question about last quarter's expenses or utilization (or any other measure you're capturing) can drill down into the data (assuming they have sufficient permission), separate it, ask questions, and find answers. Information can be broken out in countless ways, looked at microscopically or as a big picture. What used to be an unmanageably huge report is now interactive in the best sense of the word. Managers or members of departments most effected by particular trends are able to personally pull apart data and ask the questions that only they would know to ask.
The healthcare field demands fast, creative solutions. If your BI environment doesn't have a high level of usability and flexibility, it might be time to reevaluate the tools you're using. Remember, having data isn't enough. Visibility is key -- getting the right eyeballs in front of the data -- as is having data available in ways that are understandable, accessible, and ready to answer your questions.
David Cole is the chief technology officer (CTO) at Claraview, where he focuses on building out the company's strategic service offerings and solutions. He's the lead architect of Claraview's Web Analytics solution. His specialties include big data, Hadoop, data-driven marketing, customer data integration, and data warehouse design. David is a subject matter expert in the field of data management and big data, cited frequently in industry publications. He holds a Bachelor of Science degree in Engineering from Cornell University.