Taking Cat Map Tools For A Test Drive

Holborn's literature describes its “Eye In The Sky” tool as a “client-driven, real-time dynamic catastrophe analysis and management tool.”

But the words come to life when Chief Technical Officer Frank Pierson sweeps a mouse across his pad, simultaneously drawing a line across Harris County, Texas, on his computer screen, effectively turning his cursor into the path of a simulated hurricane.

Similarly, Guy Carpenter's description of its “i-aXs” tool as a “customizable, Web-enabled platform with fully integrated mapping capabilities” has more impact when Managing Director Lara Mowery remotely clicks a drop-down menu on your computer to show a satellite image of an insured property before Hurricane Katrina with a superimposed radar damage analysis for the location available days after landfall.

Below, we give details from the demonstrations provided by both firms for NU early this year to help readers better understand some of the features of each tool.

o Holborn's Eye In The Sky (EITS):

As Mr. Pierson began his demonstration of Holborn's “Eye In The Sky,” his computer screen displayed a U.S. map covered with brightly colored dots. There were 81-thousand-plus dots, he reported, noting that each dot represented a policy location and each color stood for a different type of business insured by the client whose policies we were seeing.

By clicking a zoom button on the product's toolbar and then clicking and dragging his mouse to draw a rectangular box over a section of Texas, Mr. Pierson allowed the tool to zero in on Harris County.

“It would be a little difficult to do an analysis on that [nationwide] level,” he said. EITS “is a way to visualize your portfolio and then to analyze it. Visualizing it isn't really enough. You have to analyze it,” he said as he zoomed in on the county data.

As he zoomed in closer and closer, not only did the proximity of individual risks to each other become clearer, but so did geographic details, like roads, state and county boundaries, and river details.

These details started to fill up in a way familiar to users of popular Internet mapping tools for driving directions–and with good reason. Mr. Pierson recalled that the impetus for EITS came last year when a broker was visiting a client to help the customer think about understanding its catastrophe model. They were playing around with Google Maps at the same time–sparking the idea to put the two things together.

What started off as vaporware for this one client evolved over the course of the year, as Holborn periodically demoed the product to others–and added features and changed the look every few weeks in response to their feedback, he said.

Initially, “we didn't know who all the users were going to be. We also didn't know what the final product was going to be, exactly,” he said, highlighting some positive byproducts–the need to make it simple enough for nontechnical users and to let them create the reports they want to see.

Mr. Pierson, a casualty actuary, proudly reported that the tool can be demonstrated to clients by the least numbers-oriented brokers in his firm. “They're sales people, not high-powered actuaries,” he said, attesting to the fact that intuition can guide all types of users to find their way to information they need.

Demonstrating, however, that detailed numbers are also a mouse click away, Mr. Pierson clicked on the words “Quick Statistics” appearing to the left of the map to reveal a numerical summary of the dots displayed. The summary included the number of policies in the area around Harris County being shown, along with the total insured value, gross premiums, average annual catastrophe loss (AAL) and deductible. (AALs are calculated from RMS, AIR or EQECAT models. Users can select among the three or compare them.)

Like the map, Mr. Pierson demonstrated how users can zoom in–or “drill down” on this overview data as well, by clicking tabs to display the same statistics by state, county and ZIP code for property, auto or workers' compensation policies.

Displaying a long list of policies under the ZIP code tab, Mr. Pierson simply clicked on the commercial property premium column to sort the policies from highest to lowest premium. Then he clicked the top policy number displayed to pull up a satellite map image of the policy location along with detailed data such as insured name and address, type of business, and construction code.

Spying another policy in the vicinity on the satellite map, he clicked on the dot for that policy to get the same details.

He pulled up similar types of summaries and dotted maps by drawing boxes around major counties and highways and around earthquake fault lines and other indicators of catastrophic peril. He displayed fault lines, terrorism targets and historic hurricane storm tracks on the maps by clicking on the toolbar icon for a library of major perils, and was also able to analyze all policies within a certain distance from the Florida coast using something called a policy filter.

As he clicked on headings in summary data tables to instruct EITS to sort policies, or dragged his mouse to define regions for study, Mr. Pierson observed that EITS gives its users “pockets of information,” prompting them to think about their risks and what additional details they want to understand those risks better.

“We've had clients ask, 'Is there a list of reports I can get?'” The answer is no, not really, he said, “because we don't really know what you want to look at. The idea is for you to look at your portfolio and figure out what it is you want to do.”

During the demonstration, Mr. Pierson rarely touched his keyboard. In one instance, it was to type the digits 250000 into a policy filter dialog box, indicating that the only property policies he wanted to analyze were those with total insured values greater than $250,000.

“It's all mouse-driven, so it's easy” to use, he said, noting that while users need not be experienced in spreadsheet mechanics, EITS does allow them to download requested data into Excel for more analysis.

Prompting the observer for her address at one point, he also typed that in, revealing EITS' ability to calculate the impact of adding one or several risks to a portfolio's probable maximum loss.

While that feature may be particularly important to an underwriter, Mr. Pierson said EITS was designed without a target user in mind.

“Some clients want to put it on the underwriter's desk. Others have said, 'No, my underwriting is done by my underwriting rules and this is going to be for management oversight,'” he said, putting more users to date in the latter group of chief risk officers, chief underwriting officers and other high-level managers.

He reported that reinsurers could also use EITS. Instead of defining the colors of dots to be auto dealers, lumber yards and other business types, yellow could be XYZ Mutual and green could be ABC Mutual. “We don't prescribe what data fields have to be. The system is dumb that way and smart that way, depending on how you look at it.”

He said one feature of particular interest to many customers allows them to design their own catastrophes and view their impact. By clicking a peril tool on the toolbar, he called up a dialog box prompting him to choose the category of his storm (ranging from a tropical depression to a Category 5 hurricane), its intensity (weak to strong), its speed (up to 200 mph) and even damageability factors by construction type.

After clicking on the word “simulate” and drawing a storm path by sliding his mouse, EITS displayed an array of concentric circles representing varying levels of maximum sustained winds around Houston, which the tool effectively “whacked with a Force 5 hurricane.”

While many policy location dots sat within the circles, his particular storm caused only $3.7 million on a gross basis, because of wind exclusions on many policies.

But for bigger totals, he said, a next step would be to view the reinsurance impact. A “Partner Portal,” which summarizes reinsurance by layer and by individual reinsurer within layer, is used to tally the impact, he said, while clicking a reinsurance tab to link to the portal.

o Guy Carpenter's i-aXs:

In an online presentation of “i-aXs” that Guy Carpenter shows to clients, the image of a data monster–a huge creature covered in scraps of paper–walks across the screen. As the monster pauses to put a light bulb into a socket on the ceiling of a previously dark room, the words, “Now your data is more useful than ever before” appear.

When Lara Mowery remotely accessed NU's computers to demonstrate i-aXs earlier this year, she didn't show the data monster, but the idea of organizing mountains of data was the focus of her remarks.

“I like to think of i-aXs as a very large data warehouse with very robust report-writing capabilities [and] integrated mapping and satellite imagery built onto the platform,” she said, going on to stress that organizing the information in the warehouse is what distinguishes her product.

“A lot of the analyses built into i-aXs are things that everybody does–that a lot of insurance companies do themselves and that Guy Carpenter had always done outside i-aXs. The difference is how it's delivered, how it's organized,” she said.

“The information is organized so you can see it the way that you want to see it. That makes it usable, as opposed to painful to get to an answer,” she said, noting that hundreds of pre-built reports–both reports that are available to all users and special reports customized for individual users–reside in the system.

“They're there for you,” she said, going on to show how to access the reports by clicking on them from a customized home page–i-aXs' launch point.

“Each username and login has a unique homepage. Just as you would sign up for content on MyMSN or MyYahoo and list 10 things you want to see when you log in, this is the same way,” she said.

Appearing on the screen were six boxes of numbers and maps with headings like “Limit & Deductible By State,” “Alerts,” “Accumulator” and “AAL Reports.” Each box represents a category of reports, she said.

“If you're interested in…how high up your policies are attaching [and] what does the [limits] profile of your business look like, you would sign up for 'Limit & Deductible' reports,” she said.

Each user picks report categories that are relevant to their day-to-day activities, she said. The CEO will have a different view from the underwriting manager, for example, because they can pick completely different content, but all of the reports are being generated off the same database.

“You're not stuck with what you've picked to put on your homepage,” however, she added.

For example, a user might have signed up for “Accumulator Reports” in five regions, which help them identify and visualize exposure concentrations on maps in selected regions–like accumulations in Florida around sinkholes. But if New York wasn't a hot spot for the company before, a user can add the state by selecting it from a catalog of all available reports, she said.

That new report will run in anywhere from five-to-45 seconds, complete with the mapping, she said, explaining that if the user wants to continue to see this information daily, they simply click “subscribe” to add it the homepage going forward.

Every report selected for the homepage is automatically updated when new information is uploaded to the database, she added.

Ms. Mowery went on to demo other key features. “Data Miner,” she said, is a core piece of the platform that contains the report-writing tool with the integrated mapping and satellite imagery, as she displayed a slide with all three.

The report grid listed total insured values (TIVs), average TIVs, deductibles, and the number of risks by state and construction type. On a corresponding map, different shadings represented different ranges of TIV amounts. “Once you're in the map you can pick what you want to shade by,” she said–noting, for example, that number of risks could be selected rather than TIV.

As in EITS, Ms. Mowery demonstrated that i-aXs users could drill down from a state all the way to individual location level within reports and maps, and also turn on pre-loaded hazard layers of historical hurricane tracks, known fault lines and other perils, although different mechanics get to these features within the two products.

Describing i-aXs' “Accumulator” platform, Ms. Mowery said it evolved from internal work Guy Carpenter did after 9/11, when clients wanted to better understand their accumulations of risk near terrorism targets. She demonstrated the feature by having i-aXs draw shaded concentric circles around Seattle's Space Needle, which indicated increasing ranges of TIVs.

“You can run a report that tells how much exposure you have within one mile of every one of the locations in your portfolio,” she said, adding that i-aXs turns out accumulation reports–complete with maps–within 30 seconds.

Before i-aXs, a best-case scenario would put this at a three-to-four day exercise; worst case was two weeks, she noted.

Ms. Mowery went on to show other i-aXs features, including:

o Policy Ranking–a downward sloping curve on a graph that shows which individual policies contribute the most to a firm's cat modeling results.

o Gradient analysis–a shaded map indicating areas of high exposure concentrations and high propensity to catastrophe loss, such as hurricanes. Lighter zones are the areas where an insurer can continue to add business without negatively impacting its cat-modeling results, she said.

She also highlighted i-aXs' ability to evaluate the impact of an event as it is happening as a key benefit, noting that built-in streaming real-time weather information makes that possible. A map she displayed as she spoke showed blue dots for individual policy locations and boxes in different shades of red–with one dark shade indicating, for example, an 85 percent or better chance of at least three-quarter-inch hailstones.

Referring to this feature as “RealCat,” she said “this is of great interest to claims professionals fielding phone calls from senior management,” going on to describe a companion feature known as “i-Map.”

i-Map allows a user to enter in a location, to view all available satellite imagery of that location and “to load over that hazard information that tells what happened during that event.”

She explained that three days after Hurricane Katrina, Guy Carpenter combined pre-event satellite imagery with a radar analysis of flood and wind damage locations provided by the Federal Emergency Management Agency (at the time putting these images on the reinsurance broker's public Web site, because the full i-aXs product was in development).

A red layer covering parts of the satellite photo revealed where the damage had occurred based on the FEMA analysis.

In addition, Guy Carpenter requested post-event images from a satellite data provider, making this available in less than two weeks, she said.

Additional information–such as elevation of the location, estimated storm surge height and wind speed provided by i-aXs–is also useful in assessing damage, she said, displaying a post-event image of a seemingly intact hotel and casino complex.

“There was actually a 15-to-20-foot surge at that location,” she said, turning on a data layer, which revealed something the overhead image did not–that the bottom two floors of the building were totally wiped out.