LAS VEGAS -- In 2007, Alfa Insurance, an insurer that operates mainly in Alabama, Mississippi and Georgia, had no data analytics infrastructure in place. A year later it had deployed an operational data warehouse. By 2010 it was running analytics software on top of the warehouse and basing underwriting decisions on the results of data analysis applications.
It allows you to compete with the big boys if you do it right, and it's not beyond your reach.
vice president of business development, Alfa Insurance
Speaking at the 2014 TDWI BI Executive Summit here, Mike Rowell, vice president of business development at Alfa, said he initially felt lost when it came to implementing a data analysis infrastructure from scratch and getting to a point where the organization could use analytics to drive decision making. But he added that his experience proves it's possible to get sophisticated analytical capabilities up and running in a relatively short period of time.
Ultimately, the cultural challenges proved to be more difficult to solve than the technical ones at Alfa. Rowell said the Montgomery, Ala., company had to deal with the usual problems of siloed systems, inaccessible data and data quality issues. But the hardest part was getting the organization to embrace data-driven decision making based on a unified set of information. The prevailing view across the insurer's 17 business units, he said, was, "If it's not my report, I don't trust it."
A big part of changing such views is getting the right people on board, according to Rowell. He said the idea to make the company more data-driven came from upper-level management, particularly Alfa's chief operating officer. Getting executives to support an analytics initiative helps immensely, he said. "If you don't have that endorsement, do everything you can to get it."
Not failing is not an option
Next, you need to build an analytics team. For Rowell, it was important to find people who could tell him about times they had failed and explain what they learned from the experience. Analytics projects inevitably encounter challenges and don't always work out as planned, he said, adding that someone who has never experienced failure is "not going to be able to deal with that."
Finally, you have to demonstrate the business benefits of an analytics service. And at Alfa, it really is a service. Simply putting in a place an analytics infrastructure that could assist business executives in making decisions wasn't enough, Rowell said. "One thing we learned is that if you build it, they will not come," he said. "Managers are busy. They don't have time to play with data, at least in my experience."
Rowell said proving the value of analytics is necessary to get business units to adopt new ways of doing things -- and new business strategies. For example, Alfa saw its overall loss ratio -- the difference between claims paid out and premiums collected -- drop by nine percentage points from 2010 to 2013 because the analytics capabilities enabled it to focus marketing campaigns on higher-value customers. The process wasn't easy -- Alfa dropped more than 10% of its property insurance customers. But the level of improvement in the loss ratio "is huge," Rowell said.
From a technical perspective, organizations need to be strategic when they begin implementing analytics initiatives. Peter Mueller, head of the global business analytics program at Lonza Pharma & Biotech, a pharmaceutical maker that's part of Switzerland-based Lonza Group Ltd., said during another session at the summit that analytics efforts should be aligned with specific business goals from the start.
Analytics infrastructure requires clear thinking
Lonza, which began working toward a more unified approach to data management and analytics in 2011, initially fell into the trap of trying to just get something in place to start collecting data as quickly as possible, Mueller said. But he added that without a clear idea of what you want to get from an analytics system, you're likely going to end up with underutilized technology.
Mueller said Lonza is looking to integrate the business analytics program into all of the pharmaceutical unit's operations by 2015. A key part of the deployment strategy, he noted, is to try to "get to the bottom of the silos" and uncover the original planned uses of the data in them. "Somewhere along the line, someone developed a business case for the collection of that data," Mueller said. And in his view, the closer the analytics program comes to satisfying the needs of the data creators, the quicker it will be embraced and adopted by business users.
The challenge of implementing a data warehouse and analytics architecture from scratch can seem daunting, but the business opportunities for those that are successful can be great. Rowell said his midsize insurance company now has data analytics systems that are comparable to the ones at much larger insurers operating nationally. He sees the analytics infrastructure leveling the playing field.
"It allows you to compete with the big boys if you do it right, and it's not beyond your reach," Rowell said.
Dig deeper on Business intelligence data mining
Ed Burns, Site Editor asks:
Where do you think a business should start when implementing an analytics architecture?
2 ResponsesJoin the Discussion