The saying goes: "If you want something done right, do it yourself." And, when it comes to sales and marketing analysis, risk management and financial reporting, insurance analysts, managers and executives are taking matters into their own hands-thanks to software that gives them immediate access to corporate data.Instead of waiting in line for an IT department to generate static (often paper) management reports, more and more insurance companies are employing business intelligence tools to interact directly with the data they need for planning and decision-making.
With Internet communications and user-friendly interfaces, these tools enable non-technical personnel to "slice and dice" information, and drill down several layers into items in a report for a highly granular level of detail.
With business intelligence tools, reports are very different from the way they were when generated by management information systems in the past, says Uri Taiber, president and CEO of InsFocus Systems Ltd., an Israel-based business intelligence software firm.
"Five years ago, I was an insurance executive and I would have to define some kind of management report I needed," he says. "I would set the parameters, send the request (to the IT group), then wait to get the printed output the next morning. If I didn't define the parameters properly, I would get some garbage back-and I'd have to sit down and define my request again. It was really counterproductive."
Now with business intelligence software, managers can get reports much more easily and almost immediately, says Taiber. "Within 10 or 20 seconds, you can see whether your question is answered or not."
Indeed, over the past several years, insurers, especially the larger ones, have been implementing business intelligence and proving its value. According to Framingham, Mass.-based IDC, the market across industries for business intelligence software is worth more than $7 billion worldwide and it's expected to double by 2006.
"Data mastery," which includes business intelligence, ranked among the top three projects that delivered the highest return on investment in a CIO/CTO survey conducted last year by Celent Communications Inc., Boston.
The same study revealed that business intelligence projects were a "high priority" or were already "in progress" at 42% of property/casualty firms, and 63% of life insurers surveyed.
Many of those firms have been large carriers, but smaller regional insurers are beginning to buy into business intelligence as well. "There's a big disparity in uptake between larger insurance companies and the smaller guys," says Leo Tucker, director of banking solutions at Cognos Inc., Ottawa, Ont.
The top 100 carriers typically have a business intelligence strategy and have had the products for anywhere between three and 10 years. But for smaller carriers, he says, it's still not uncommon to see companies investing in business intelligence for the first time.
In the right hands
United Heritage Life Insurance Co. is a regional insurer that invested in business intelligence sooner than most carriers of its size. Four years ago, the Meridian, Idaho-based company implemented a business intelligence tool from ProClarity Corp., Boise, Idaho.
"We needed to get data into the hands of the users without having to go through the IT group," says Mick Ware, vice president of information technology at United Heritage Life. "That was the main reason: Just getting IT out of the loop of generating and regenerating reports."
Approximately 25 United Heritage decision-makers now use the business intelligence tool regularly, and about 10 others use it intermittently for marketing, underwriting, actuarial and new business analysis.
Only one person in the IT department is required--and only on a part-time basis--to develop and maintain the multidimensional databases, called "cubes," that provide end users with specific, updated information culled from the company's administration systems.
For instance, agents and managers can now see sales and production data including agent sales, persistency rates and paid-to-written ratios. Administrative personnel can track key performance indicators any time, from any desktop, including written, delivered and terminated premium for all products broken down by agent, by the type of contract they have with the company, and by the status of the policy.
"Marketing managers can run a general production report for a region for a specific time period," Ware explains. "They'll see a graph that shows them production numbers for certain lines of business. And if they see an anomaly, they can click on the graph to go to another graph that shows a specific time period, or a specific product, for example. Then, they can actually drill down to a record level of detail that makes up that graph."
Data that pays
Ware doesn't know how much time or money United Heritage Life has saved using this tool, but he is certain that marketing managers, underwriters and actuaries don't hesitate to use these reports whereas before they may have hesitated to ask IT to generate them.
"Plus, we know in IT that we don't have to write reports and go around and around with users like we used to do," he says.
Corona Direct, a Belgium-based insurer, knows exactly how much it is saving since implementing business intelligence software to identify prospects for funeral expense insurance. "We've been able to mail 30% fewer mailers, reducing our costs by 30%, and still get the same results," says Philippe Neyt, commercial director, at Corona Direct.
In the past, the company's acquisition costs for this particular product exceeded its first-year revenues. But two years ago, Corona Direct replaced its hit-or-miss process of targeting prospects based on human experience with an automated predictive tool from Chicago-based SPSS Inc.
Based on Corona Direct's historical data on the actual behavior of funeral expense insurance prospects, the predictive tool identifies people who are likely to buy the product. The software then performs a profit-cost analysis, balancing growth targets against profit margins. "The probability that someone will respond to a mailing for this product is extremely low--0.7%," says Neyt. "So we had to find a way to reduce the costs for the same amount of sales or quit selling the product through this channel."
Ultimately, Corona Direct increased sales efficiency by 36% and recouped its investment in the software within a year. Furthermore, for the first time, the company is making money using the data in its data warehouse, according to Neyt.
"We were doing all kinds of analysis with our data before, but this was the first time we were able to make decisions with it that make more money for the company," he says. "That's what software should do."
Many insurers, such as Corona Direct, realize they need business intelligence tools when they grow to the point where their data becomes unwieldy, sources say.
"You can make calculations with the human brain and standard software on simple criteria, such as whether a prospect is male or female, or lives in the northern or the southern part of the country," Neyt notes. "But beyond about 10 criteria, and whether or not one criterion influences another, the analysis becomes too complex."
In fact, if companies are small enough to get information when they need it, they probably haven't needed business intelligence, says Cognos' Tucker. "But as they grow, companies typically have multiple data sources and more information than they know what to do with. So that's when they might say, 'Let's bring business intelligence in to help make sense of all this data and turn it into clear and concise information,'" he says.
In addition, Tucker notes, because business intelligence tools access data across multiple data sources, they serve as a central collection point for information across systems-systems that proliferate as a company grows.
Single source of truth
"We needed a better understanding of our customers and our business, and a single source of truth so we would have consistent and accurate information readily available to anyone in the organization who requires it," says Stan Pachura, vice president of strategic technology at PMI Mortgage Insurance Co., a Walnut Creek, Calif.-based mortgage insurer.
Last year, PMI began implementing a suite of Cognos business intelligence tools, which supports corporate performance activities from harvesting data all the way through planning and achieving the company's strategy.
"We use Cognos' ETL (extract, translate and load) tools to gather data from various sources, then translate that data into definitions we've identified as the data enters the data warehouse," Pachura explains. "We also use Cognos tools to build data marts, or cubes, that are specific to certain part of business-daily sales information, or claims information, or information specific to the general ledger."
Within each of those cubes, PMI decision-makers can slice and dice the data and look at it from many different perspectives, such as geography, customer, or time, he explains. The company also uses a reporting tool that enables users to create ad-hoc queries, and an enterprise planning tool that ties business intelligence back into the company's planning, budgeting and forecasting activities.
"PMI is visionary in its adoption of the overall corporate performance management concept--of which business intelligence is a solid component," says Cognos' Tucker. The other components include developing a strategic plan and budget, and monitoring business performance against the plan-typically by using alert mechanisms and scorecards.
Classic business intelligence tools enable companies to get a better understanding of what's happening in their operations, Tucker says. "If you're monitoring some part of the business and you need more information, you use business intelligence tools to get a better understanding of what's happening. Based on what you learn, you feed that back into your planning process to close the loop."
The leading edge
Taking information from a business intelligence system back into operational systems--referred to as "in-line analytics"--is the leading edge in this market, notes Anant Jhingran, director of business intelligence in the Silicon Valley Lab Software Group for IBM Corp., Armonk, N.Y.
In fact, in July IBM acquired Alphablox Corp., a privately held company based in Mountain View, Calif., to provide this capability as part of IBM's data management and on demand computing initiatives.
Another recent trend in the business intelligence arena is that larger carriers are consolidating platforms, according to Cognos' Tucker. "Many larger insurers have bought business intelligence products for departmental use," he says. "Now, they are standardizing reporting and analysis across the enterprise-on one or maybe two products instead of four, five or 10."
Highmark Inc., a Pittsburgh-based health insurer, is one such firm. In 1998, the company decided to consolidate its analytical resources scattered across the organization into one large "informatics" group.
"Once the group formed, we began to make decisions about technology-both hardware and software," says Richard Pro, vice president, healthcare informatics research and analysis at Highmark. "From 1998 to 2001, we made some hardware upgrades and we moved our analytics to the Teradata platform, which enabled us to do a lot more complex modeling."
At the same time, the informatics group developed its modeling techniques, which include a combination of off-the-shelf packages and homegrown software. "The reason we invested in this manner is that the industry has become much more complex," Pro says.
"It's no longer good enough to do analysis in hindsight. We have to be able to do predictive modeling. We have to understand the risks in our population from a health standpoint so we can improve our medical management of those risks."
To that end, the 100-person informatics group works with the sales staff to assess risk pools and peer comparisons associated with their employer accounts. The group ties those risks back to specific clinical issues at the account level to help employer groups determine which Highmark programs may help them manage their costs.
"For one account, our diabetes management program may be critical," says Pro. "For another, it may be our coronary artery program."
Highmark's informatics group also provides analyses for senior managers to simulate the effects of strategic decisions, and for internal managers to identify high-risk individuals for clinical and financial management of chronic diseases.
"Employers are asking for changes in benefit structures," Pro says. "And these days it can be very complex for executives to evaluate what their product line should contain--what types of products and what benefit structures. So it's very valuable to simulate what will happen as a result of those changes before we actually make them."
It's also very valuable to ensure that Highmark is devoting resources to managing high-risk cases. To target the right people for chronic disease management, the group developed a stratification process for members using data mining and predictive modeling tools from Cary, N.C.-based SAS and software from Boston-based DxCG Inc.
Before implementing the technique, the company used simple clinical algorithms that often led to false positives. For instance, in an early test, the team identified nearly 35,000 potentially diabetic people, and 8,000 of those were false positives. Now, for the same 35,000 members, only a handful would be falsely identified, says Pro.
"We were missing people who had conditions and not getting them into the program. Meanwhile, we were having our condition-management vendors contact a lot of patients, only to find that they didn't have the condition," he says.
Reducing false positives to nearly zero has produced significant cost savings, Pro adds, because now the company is investing resources in high-risk people, and not on those who don't need intervention.
The company is also reducing costs associated with fraud. Fraud in the health care industry is a multibillion-dollar problem, Pro notes. "If we're experiencing as much fraud as we see in the national average, it could be $100 million problem for us."
Using a hybrid analytical technique based on SAS tools and neural network software from NeuroDimension Inc., Gainesville, Fla., Highmark has improved its return on investment for its fraud program from 3:1 to 9:1. "For every $1 million we spend on fraud detection, we save $9 million," says Pro.
And, making or saving money is what business intelligence is all about, according to Corona Direct's Neyt. "There's a lot of software on the market that makes all kinds of magnificent analysis and fantastic graphs and tables, but it's not meant to go into action-to sell more products or to reduce costs. The main reason we chose SPSS was that it is action-driven software," he says.
Zurich NA Uses Intelligence To Deliver On Its Claim
When you profess to be the lowest loss-cost provider, you had better deliver on that claim. Zurich NA is carrying out that value proposition by providing Web-based business intelligence to its large account customers-to help those customers proactively manage their risk and lower their loss costs.
Launched in 1998, the extranet, called RiskIntelligence, provides approximately 6,000 Zurich NA's customers-primarily those with a high frequency of workers' compensation claims-with more than 100 standard reports, including data by customer location, type of injury, product line, and size.
"They can see each claim they have with Zurich, they can drill down to adjusters' notes, and they can view their financials," says Marcia Engel, manager of customer information services at the Schaumburg, Ill.-based insurance company.
The Web-based tool also provides Zurich NA's customers with trend data, such as slips and falls that occurred with Zurich as their insurance provider as well as with previous carriers. Benchmarking data is also available, which compares Zurich customers to industry averages for specific losses.
"Right now, we're developing data visualization, which will enable us to take a full 30-page report and bring the most serious loss trends up to a single page so customers can drill into the report from there," says Frank Colletti, director of technology services at Zurich NA.
The system, powered by WebIntelligence from San Jose, Calif.-based Business Objects Inc., accesses claims data in Zurich's data warehouse. "Our data warehouse was being used by our claims offices to track trends at the office level, and we took that same data and built a sister database that is secured for our customers to access through the extranet," says Colletti.
But RiskIntelligence is far more than a data dump, he adds. "We've had cases where they've identified loss trends that have prompted risk engineering to be done," he notes. Before implementing RiskIntelligence, it could take months before a loss trend would become evident. "Now, the data is updated daily, so customers can apply risk engineering and save money over time."
