14 Jan Visions with Matt Redlon: Seeing business intelligence clearly
Madison, Wis. – Don’t tell Matt Redlon that business intelligence is oversold.
Redlon, chief information officer of the Eden Prairie-based Clario Analytics (formerly Decision Intelligence), believes that when it comes to exploring the inner workings of business at its broadest levels, business organizations are wise to leverage tools that mine their data.
To Redlon, who was in Madison recently for a Wisconsin Data Management program, getting some answers to drive business performance management is a very real business need – especially when business clients are struggling with fundamentals like getting to know their customers and understanding the very nature of their enterprises.
“In a lot of ways, they create windows into your data warehouse,” he said of the aforementioned tools. “The provide a view, and at a high level they do things that are as simple in concept as an Excel pivot table on your underlying data.
“If you think of it that way, I don’t think it’s oversold.”
Redlon, who holds a bachelor’s degree from Northern Michigan University and an MBA from the University of Minnesota, played a role in designing Clario’s web-based analytics platform. He believes web-based analytics has its advantages, but acknowledges that software-as-a-service – delivering an application over a web browser – can be over-hyped or misunderstood.
Still, SaaS and other web-based applications make a lot of sense to him, especially in the current economic environment. “The applications that make the most sense in that framework are those that take advantage of the cost economies of multi-tenancy,” he said. “The fact is that you get those economies of scale for storage and computing power. To throw out another buzz phrase, those economies are typically associated with applications that make heavy use of the cloud, right?”
By “cloud,” Redlon is referring to a cloud platform such as Amazon’s Web Services, Google’s application engine, and the new Windows Azure platform. He said it’s about looking at the server and storage cloud now being made available in what is described as platform-as-a-service, where servers and storage are perfected internally and exposed externally for outside users on a utility basis.
Web applications that make the most sense, he added, are ones that can take true advantage of those platforms, which offer tremendous economies of scale. “In our particular case, data mining and data analytics is very spiky,” Redlon explained, “and you might not go in and run large-scale analyses every day. You might do it very infrequently. Being able to access that broad capability that might be needed on an infrequent basis, and building out the ability to scale that usage internally probably does not make sense from a cost effectiveness perspective versus using it from an outside vendor.
“Web distribution is the ideal way to access that scalable computational power from outside your own firewalls.”
Asked if it’s typical for organizations to build the data warehouse internally and then go outside for the extraction tools, Redlon said it’s fairly atypical historically. Most organizations are building their enterprise data warehouse and then looking to bring in packaged business intelligence applications and appliances, but he said the historical pattern involves an application server and then a tool that creates local “cubes” and accesses those cubes via an Internet dispersed browser.
“In the case of our application, it’s extracts from the data warehouse that are purely sent to the application for the purpose of data mining,” he explained. “In the broader sense of BI, I think there are going to be significant challenges in the near term for enterprises to deploy an off-premise or a web-based business intelligence tool because of the practicality of remote data access.”
That isn’t the only challenge Redlon sees, however. There are both security and simple transport costs if an organization ships large quantities for standard reporting outside of its firewall. “If you take a really long time horizon as your view, and you think about books like Nick Carr’s “The Big Switch,” which talks about the move toward utility computing, in the longer term the cost economies of clouds are going to lead toward some organizations actually building their warehouses external to their organization. But that’s a long time horizon before that is an accepted thing by the enterprise.”
BI in the healthcare space
One of Redlon’s career stops was United Health Group, which has applied business intelligence to good effect. While Redlon praised United Health’s work in this area, he does not believe it’s a given that healthcare organizations will aggressively move to capture data once electronic medical records and the healthcare information technology systems that feed them data are more fully adopted.
“I think it’s going to take more cajoling,” he predicted “First off, there have been a lot of initiatives toward the adoption of electronic medical records that have almost universally failed. That’s probably too broad of a statement, but there has been very limited success in rolling those out.”
There also have been some stand-alone success stories, particularly in what Redlon calls closed systems. He cited Kaiser, which controls multiple points – membership, providers, and payers. In such cases, the EMR and, more broadly, its business intelligence uses have functioned pretty well.
“But when you’re talking about an open system, with many different providers and many different payer organizations and where membership is interacting with all of those, with the cohesiveness that’s required for true business intelligence and best practices and evidence-based medicine practices, all of those struggle a lot more outside of a closed system,” he said.
Given the interest in adopting the multiple HIT systems that can feed data into EMRs, business analytics might not be that high a priority until all major HIT pieces are in place. The situation reminds Redlon of the old IBM commercial with the mythical “universal business adapter.” It has plugs coming out of it in all directions, and the staff members all are smiling about how it connects all of the company’s infrastructure.
“That’s the parallel in my mind,” Redlon said. “So many of these payers and provider organizations have legacy systems or new systems, and there is no standard interchange for all of these systems. It opens up a huge opportunity for integrators who are looking at this as an opportunity, but I think that’s going to be a significant barrier.”
That hasn’t stopped some healthcare organizations from using BI tools to examine claims data, and to look at cost, utilization, and drug efficacy to make significant changes based on quantitative evidence.
Redlon said United Health Group established the United Resource Network around provider outcomes for major operations such as heart bypass surgeries. It looked at the performance of a provider of heart bypass surgery at a facility versus another facility based on the evidence-based business intelligence culled from claims data. The network tried to compare outcomes and follow-up costs, and established the practice of sending patients to preferred facilities with better outcomes.
According to Redlon, they would even assume financial risk and underwrite from other organizations the risk of taking those patients and sending them to the preferred providers because they were convinced that the outcomes and the resulting lower costs were real.
Even with the remaining obstacles, it never fails to amaze Redlon that BI is not adopted to a greater extent in healthcare settings.
“Absolutely,” he said.