Business Intelligence Network - The Vision for BI and Beyond

As in most industries, the data warehouse of a Lloyd’s of London insurance company is populated with the metrics and dimensions that describe the insurer’s own particular, specialised world. The nouns and numbers of insurance are immediately recognisable to their users, but almost entirely unrecognisable, and often wilfully opaque, to anyone else. The London market for insurance, as transacted historically – and very profitably – in and around the Square Mile, has its own dialect, stranger again than the broader insurance lingo of the wider industry. It deploys a barrage of specialised metrics, served up by its various systems for many different users, for many different purposes. Read through the data dictionary or shared business vocabulary of any Lloyd’s insurer’s business intelligence (BI) system, and you will certainly find an abundance of arcane language, some of it seemingly salvaged from Doctor Johnson’s original Dictionary, but now pressed into service as an attribute of a dimensional table in a Kimball star schema on a SQL Server 2005 MI system with ProClarity and a beta of Office 12, just for luck. 

In many ways, the ancient and the modern worlds are drawn together in the business of insurance in the City. The data speaks of cargo and specie, livestock and legal expenses, right alongside space risks, nuclear property damage and UK standalone terrorism. Every age adds its own lexicon to be modelled and measured and scrutinised. If the numbers show that tanneries and textiles are no longer the big ticket businesses, why not expand into bio-chem? Or astrophysics? The business of determining exactly how well each industry (or “interest”) is performing relative to any other forms a classic insurance/business intelligence problem. A careful insurer will hedge their portfolio, selecting a wide range of different risks – it’s all about risk – to offset the devastation of hurricanes against the solidity of ocean-going tankers that, year after year, stubbornly refuse to sink and spill their precious cargoes over miles of coast. Apart from when they do, of course.  

To complicate matters further, the market frequently operates with a “third-party” channel, a traditional, even gentlemanly arrangement, which, nonetheless, requires the utmost business intelligence scrutiny. Brokers bring business from Mr. Onassis and his cohorts to the underwriters of the London market, along the way taking their own, hopefully modest percentage – a little for a motor fleet in Ipswich, a great deal more for a fleet of ferries in Indonesia. Are the brokers bringing us “good” risks or terrible ones? Are their fees competitive? Do they pass on their clients’ payments in double quick time? Or are we to send the boys round every month? Hopefully, business intelligence can help with the answers. 

Among the whirlwind of proprietary and esoteric terms – many of which we will address later – sit those two familiar, almost household metrics that no insurer can live without and which form the basis of their entire business model: premiums and claims. In our daily lives, we all have a little experience with these two rogues. But, like the Eskimos and their legendary fondness for naming snow, insurers are caught in a blizzard of epithets for the stuff that continually swirls about them. Premiums and claims figures come in a large range of flavours for what is, surely, quite a simple pair of transactions.  

Definitions of premium alone include: estimated premium income, gross written premium, gross 100 percent premium, gross net premium (yes, really), signed premium income, and many, many more. Some of these serve to satisfy the statutory and regulatory "requests" for information from the plethora of controlling bodies in the London market: Lloyd’s Corporation, the Financial Services Authority, Her Majesty’s Revenue and Customs and any number of foreign legislatures, demanding to know what insurance business is being transacted on their turf. Other definitions have insinuated themselves into the market with the tidal movement of staff and techniques from the commercial insurance companies, and from the many mergers and acquisitions that help to characterise the industry’s turbulent last twenty years.  

Premiums can be sliced up in many different fashions. One taxonomy may insist on dividing the premiums into: 1) what the broker says we’ll get, 2) what the brokers says he has paid and 3) what the broker has actually paid. Another will attempt to slice-up the premium over the course of the policy term, to determine how much has been “earned” at any particular time, while a further version will determine whether the renewal of the previous year’s policy presents less or more of a risk. Possibly the pork pie factory has had a new sprinkler system installed – generally considered to be a good hedge against catastrophe – or, maybe, it has decided to begin deep-frying its products, with all the employee/boiling oil interface problems that may subsequently result. Thus, the risk is assessed, the premium is calculated and the BI system ingests the data before sitting back, watching and waiting for the claims to come rolling in. 

Then there are the currencies. A worldwide business will accept premiums in many currencies and pay out claims in many more, depending on where in the world the risk may be. But it likes to keep its accounts in a basket of familiar “settlement” currencies, such as U.S. dollars, Euros and Yen. Is it a good idea to record the rate of exchange of every single separate transaction, and then add up all those transactions to determine the value of the portfolio? Or does this mixing of rates create a hash total, especially if viewed over an entire underwriting year? If a claim for $10,000,000 landed on the electronic doormat last year, and has since been mired in legal turmoil, do we eventually settle at the current nominal exchange rate, or the rate at the time of the claim? What are the applicable accounting standards? And what if the underwriters want to view their books in a different way than the bean-counters? 

These issues and many others are a reflection of the subtle complexities and often harsh realities of the London insurance market. A couple of disasters in quick succession can bring a lot of pain to bear on insurers, in the form of multiple, huge claims. Insurers naturally seek to cover their exposure to large risks through the mechanism of re-insurance, often with specialist re-insurers who have grown to be some of the most profitable and prominent companies in the City. This hedging hasn't always been as closely regulated either internally or by the authorities as it is now. At the end of the eighties and in the early nineties, after a series of natural disasters (remember Hurricane Andrew?), it became apparent that certain insurers had unwittingly implemented their reinsurance in a “carousel” fashion. Put simply, company A reinsured company B, company B re-insured company C and company C was, you guessed it ,a re-insurer of company A. Oops. A couple of tropical storms on the Gulf Coast and that particular house of cards was levelled, nearly bringing down the entire market in the process. It certainly contributed to the introduction of various layers of regulatory scrutiny. Needless to say, monitoring re-insurance has fallen squarely within the business intelligence remit, and remains one of the most complex problems to be solved. We’ll examine that area in future articles. 

With the advent of ever-cheaper hardware and software, along with the constant need for insurers to find and develop a competitive advantage in the fertile-but-deadly marketplace, insurance business intelligence has a long and distinguished career ahead of itself. As Doctor Johnson himself said: “Knowledge is of two kinds. We know a subject ourselves, or we know where we can find information upon it.” Let’s help the insurers raise their game and maybe tell ‘em a few things they didn’t already know.

If you found this article helpful and would like to receive the latest insights each month from Mike Ferguson and other experts featured on the Business Intelligence Network, please subscribe to the UK Business Intelligence Network Newsletter.