No one ever said guiding a distribution network to optimal performance is easy. The insurance and financial services industry's supply chain is as complex as ever, and carriers across all lines want to exploit its value.How do you know how much business you may be missing? How do you hold your distribution network accountable for increased growth and profitability? For many carriers, it becomes the classic "what to do with all that data" dilemma.

Knowing that a virtual sea of data is available about products, the distribution network, customers, prospects and even the competition, insurers are challenged with finding ways to leverage its inherent-and often obscure-knowledge for improving overall business processes and gaining market share.

And no one ever said getting an entire marketing organization to rally around a data mining initiative that would ultimately improve your company's multifaceted distribution network would be easy, either.

Rachel Alt-Simmons, director, Business Intelligence Division at Hartford Life Insurance Co., Simsbury, Conn., found this out first-hand approximately five years ago when she and her team delved into a massive business intelligence initiative that studied the performance-and potential opportunities for its distribution network. "We recognized we had an analytic need that wasn't being met," she says.

Under the umbrella of The Hartford's Financial Services Group, Hartford Life sells retail and institutional financial service products. On the retail side, variable annuity, mutual funds, 401(k) and 529 college savings plans contribute to $155 billion in retail product assets; variable annuity products represent more than $100 billion in assets. The stakes are as high as the business model is complex.

TOP-DOWN SUPPORT

"We wanted to be able to analyze profiles and preferences of existing customers and then predict buying behavior," says Alt-Simmons. "And, we wanted to make sure our marketing efforts were what they needed to be."

For the past eight years most of those marketing efforts have been managed by PLANCO, Hartford Life's Philadelphia-based wholesale subsidiary. With 300-plus wholesalers across three Hartford product lines, PLANCO markets to financial advisors and registered brokers and dealers, who in turn sell to the client. (See Business Model chart, p. 36.)

"There were some indications that if we built a model that would allow us to look at the data in certain ways, we could help PLANCO recognize and actually predict which brokers were 'endangered'-those brokers that have not sold in six months or are considered at risk of stopping their selling efforts," Alt-Simmons says. To do that, the group wanted to bring data in from the company's annuities lines and look at how brokers function.

Hartford Life's business intelligence team recognized that the challenge would entail selling the project first on the inside - essentially becoming cheerleaders for business intelligence. "We took a bottom-up approach and then pushed the idea up to the executive level, seeking a top-down blessing," she recalls. Once the team received visible support from executives, the challenges before them seemed to unfold, one by one.

TESTING THE THEORY

At the start of the multiphase project, a group of 10 users worked with Enterprise Guide, a Microsoft Windows front-end application from Cary, N.C.-based SAS that provides visual access to enterprise data sources supported by SAS and native Windows data types. As a dedicated, intuitive interface for analyzing business information stored in OLAP data cubes, it guides the user to access data across the enterprise on multiple platforms, operating systems and databases.

"We ended up being really good programmers," Alt-Simmons recalls, "but it helped us understand our need for expanded data management." SAS Account Executive John McDonald recalls that when Alt-Simmons said the team wanted to start performing data analytics, they decided to add SAS' data mining product, Enterprise Miner, to the mix.

A proof of concept, which tested the theory across all lines and all marketing-related functions, along with capability prototyping, followed, says McDonald, "but it became obvious early on that it was going to be tough to get to all the data."

Alt-Simmons admits limited access to data and data quality issues were a challenge, but so was aligning overall business strategy with business analytics.

"After we purchased Enterprise Miner, we obviously had to find an internal group willing to use it," recalls Alt-Simmons.

Even with executive-level support, the team encountered product and functional silos along the way, which led to strategy silos and even "analysis paralysis" as the group tried to push forward.

"The reality is that the internal customer can fire you at any time—especially if they don't like what you are asking of them," she says, "so you have to be able to demonstrate value."

Barriers to implementing predictive modeling and segmentation inside the company did not phase Alt-Simmons, however. Her newly defined business intelligence group soon became known as the organization's "information SWAT team," educating subgroups on the nuances of data mining, helping areas define their business needs and confirming that data mining would complement-not take over-IT efforts along the way.

When the annuities line agreed to participate, the business intelligence group began pushing the data necessary up to SAS, where the predictive models would be built.

"Annuities was a good place to start, because although data resided in different locations, such as their Oracle database, their data warehouse, third-party sources, etc.," notes McDonald, "ultimately they were able to bring all that data together and look at it - and think about what caused certain customer or broker behavior."

FINDING THE SWEET SPOT

Once internal data challenges were met, the business intelligence group approached PLANCO and told them Hartford Life sought a win-win: if they could access PLANCO's data for contribution to the data mining project, they could help PLANCO find the sweet spot-identifying where the wholesaler could have the biggest impact.

PLANCO, of course, had its own marketing campaigns already in place, and a culture that was sales-focused, not data focused, so Alt-Simmons approached carefully, explaining that PLANCO would be instrumental in helping Hartford Life's annuities business unit reach a larger organizational goal.

"This really opened the door for us in terms of changing the way our business unit was being managed," recalls Alt-Simmons.

Gaining PLANCO's support meant Hartford Life could leverage cross-organizational expertise in the process.

"Hartford Life already has strong customer service," says Alt-Simmons, "but this exercise helped us identify just who the customer is-and what this customer may be inclined to purchase in the future."

After the data is merged and aggregated into subject areas, the business intelligence team applies SAS's SEMMA Analysis Cycle (see p. 35) to look for patterns that might otherwise appear random or go undetected. The team then uses modeling techniques to determine the correct model to solve specific business problems.

For the annuities project, it is a blended decision tree/logistic regression model, confirms Alt-Simmons.

Since formally joining the predictive modeling initiative and receiving feedback from the Hartford Life's business intelligence group, PLANCO has launched a variety of marketing campaigns to determine what's most effective in getting the producers identified as "endangered" back into the fold.

FUTURE INITIATIVES

Hartford Life is adopting SAS' Enterprise BI platform to enable users to leverage the SAS routines for predictive data mining and publish them to a Web portal or, with a Microsoft Office optional add-on, work with them in PowerPoint, Word or Excel.

"This broadens the playing field to other users who want to run their own SAS routines to get results," notes McDonald.

And now that PLANCO is on board with modeling and the impact it can have, says Alt-Simmons, the "SWAT team" is scooping out next initiatives, such as how to make a casual broker a committed or top producer.

"Another idea is to focus on the 401(k) lines of business, where there are great growth opportunities," she says. "So we might ask: What types of producers are more likely to sell 401(k), based on their sales in other Hartford product lines?"

Alt-Simmons credits the initial success of the annuities' business intelligence efforts to the "start small" approach, the application of project management disciplines and a dedicated effort to monitor the project's effectiveness. A little diplomacy along the way also helped, she modestly admits.

"This is all so new for everyone," she says. "But there is less adjustment when you educate and build trust and convince them this is good stuff."

Register or login for access to this item and much more

All Digital Insurance content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access