Go Back

MIS for underwriting

CHRISTOPHER SMITH, underwriting systems manger at Janson Green, relates how his company came to purchase a management information system to meet it's underwriters need for on-screen analysis and speed of response. Sometime back in the early 1990s Janson Green, now part of Limit, the largest provider of capital to the Lloyds's market, developed a system for accumulating underwriting statistics in to a hierarchy of analysis classes. Although at the time we were happy with its performance, like most other systems of that period, it was built as a reporting suite allowing us to select and print off triangulations.

Every month since then it has spooled out thousands upon thousands of figures, most of which took the scenic route via the underwriters' desk into the recycling bin. Occasionally , someone would find the time and energy to copy the figures to a spreadsheet and, every year of course, the actuaries demanded it.

A few years ago another complimentary database was developed along similar lines, this time with a different code set and a hierarchy to give us additional analysis. We asked out IT consultants, TriSystems, to develop the reporting tools. Again, we were pleased with its performance despite the paper – whole forests of trees – which was needed to run it.

They also developed an on screen version of the system but, being resticted to "dumb terminal" monochrome screens, it was never going to be a match for a PC in the graphics department.

At about the same time we were developing a new PC desktop front end to our main insurance application package. This new system allowed the underwriting team to get a variety of on screen enquiries and printed reports at the click of an icon by using some of the same technology like Microsoft's SQL server and Excel, all glued together using Visual Basic. It could even deliver reports to a web browser.

Up against this our triangulation printing system was beginning to look somewhat inflexible and outdated. It couldn't download a triangle to Excel, of make a stab at the projection, or monitor a written premium target or print a colour development graph. Actually, it couldn't pritn a graph at all. And we needed to do all of these things day to day.

The problem was that triangle and development graphs are actually quite hard to do properly and that sort of system, call it MIS (management information system) for underwriters, would have been too consuming and expensive for us to develop. Underwriters have classically been poorly served in this area so nothing suitable was available out there to buy in.

The scenario should be perfectly familiar to anyone in the underwriting business. Most companies have some sort of underwriting statistics database. Precious few have anything on the PC desktop to do the sort of reporting from those databases which underwriters actually need – that is, in the main, triangulations and development graphs with the sort of drill-down facilities which accounts or sales departments have had for years.

Of course there has always been plenty of spreadsheet-style MIS tools around but none of them are really geared to the underwriter. Sometimes only a triangle will do. Enter TriSystems' Experience. TriSystems has gained a lot of underwriting and statistical knowledge working for a wide range of underwriters in the companies and Lloyds markets throughout the 1990s and had spotted a shortfall in underwriting desktop software.

With their Microsoft certified Windows programming team, they developed a PC package which would integrate into any underwriting statistics database, replacing the existing reporting and analysis functionality with some state of the art graphical and tabular analysis tools. By March of this year TriSystems were ready to take the project to market, and we, being a long standing client, were natural choice for their "early adopters" target list.

Our active underwriter and director, Barnabas Hurst – Bannister, despite being in favour of the system from the outset, wanted full agreement from his underwriters, thus ensuring that his team were given a tool which they would be happy to use. A three month pilot study was launched with six of our key underwriters and their assistants, to evaluate the system and the effectiveness of its integration into our existing databases. In August all the underwriters agreed they would like the product for the entire underwriting area, around 30 users in all.

I don't think that there was ever any doubt that Experience would prove to be better than the old printout system. It sounds cliched but out investment in the software has really added value to our existing databases. Experience has brought a raft of additional functionality which the underwriting team has benefited from without depreciating any prior investment in the building and maintaining the databases for the past decade.

We now have a system that can generate multi-currency development triangles with gross, reinsurance and net views, colour graph printing, plug in statistical projection modules such as craig-head curves and chain-ladder, budget versus actual comparisons, document sharing and the ability to cut and paste graphics into our office desktop suite. And all of that from a database we built in 1990.

I would be quite surprised that if every function that if offered by the Experience software is employed by every user – it is sufficiently powerful to satisfy the various demands of a broad range of users within the team, but ease of use was still hugely important. We needed to relieve the underwriters and assistants of the burden of repetitive, time consuming tasks such as statistical aggregations analysis, projection and target tracking, as well as giving them something that would help them make the most of the data they were working with.

It could have been counter-productive to have bought a piece of software which, in providing all the solutions required, placed unreasonable technical demands on the team or required vast amounts of training and a steep learning curve before we could become conversant with it. The product is really very easy to use and this is borne out by the fact that a relative novice can pick it up within a few hours. This was of considerable benefit to us.

Speed of operation is just as important as ever and we are as keen as anyone to exploit a labour saving device when we can. Using the old system, it would take a good couple of days to produce aggregate figures for a presentation to the board, or the production of reinsurance adjustments.

Much of that time would be just waiting for reports to be processed and printed, but a large and costly amount of the time would be simply the hard graft of retyping figures and doing arithmetic in the spreadsheets. With the new system, the first time we tried it, it probably took five minutes and probably four of those minutes were learning how.

It is hard to imagine it was running over the same data as the old system used to – so its quite an improvement. The big bonus with Experience is that now we only need to do anything once – as soon as we are happy with the report or graph, we can save it as a document and refresh it and reprint it next month in a matter of seconds.

So how does our IT department see the benefits?

From IT manager Ron Sterry's perspective, delivering the right functionality to the end user for a reasonable project cost was very important. Experience has not proved to be expensive to install and integrate into our databases and it does do the functions the end users want.

At Janson Green we have for some time been developing our own graphical front end to our core systems, so we are familiar with the technology that the new system uses. It works well with our strategy – it fits into the existing infrastructure, complementing the current systems and does not mandate the purchase of yet another database technology that we would be committed to support long–term. That is a big advantage of buying in this kind of system.

What then of the future?

Now the technology is in place to work with the underwriting statistics on the desktop, one of the tasks is to enhance the underlying databases to allow additional analysis by various codes such as broker, territory and underwriting unit with additional drill-down capability. This may well require a multi-dimensional "cube" approach, with which the new system will work well, or can retain and expand upon the database technology we are currently using.

Of course all of this is only now feasible with the on-screen analysis and speed of response, that the new system delivers. We are currently also looking into Experience's automatic monitoring add on, called Monitor, which will automatically track various measures such as actual verses budget, claims budgeting, cash flow projection, investment returns and so on.

The impact which the new system has had in our business area is significant as it empowers the team. Rather than having to translate the business user's idea's via the IT team, they can apply them immediately and directly into the database via the product and see the first appear first hand.

This allows us to analyse our business in a more meaningful manner, and be much more flexible and dynamic in our approach. Our ultimate goal is to have the best informed underwriting team in the London market. Our new system is a major step towards that goal.

-   Computing for Insurance