Do We Really Need Data Warehouses?
Although they are supposedly in the business of assessing and managing risk, insurance companies have historically shown an alarming tendency to throw money at expensive data management projects that never get completed, or that cannot possibly deliver the hoped-for benefits.
It is not uncommon, for example, for insurance companies to spend upwards of a half-million dollars, or much more–six to eight months of valuable time–converting all the data in their storage systems when the company makes an acquisition, confronts a regulatory change, or simply wishes to streamline a key business process.
And, because this whole data conversion process has to be repeated whenever the system must serve an important new business need, the goal of a central data repository interfacing seamlessly with all of a companys legacy systems often becomes a mirage on an ever-receding horizon.
Why do insurance companies continue to chase this elusive vision? Its obviously not because they believe in wasting time and money.
Rather, many have found themselves the prisoners of certain traditional ideas about how data storage has to be done. Many of these ideas, in fact, are implicit in the very name for the kind of data storage system that causes insurance companies such problems: data warehouse.
Think about this term, data warehouse, for a minute. Based on an analogy from the physical world, it supposes that data residing in disparate systems needs to be transported to a single location in order to be efficiently made available to end users.
But why should this be so? Why should "bytes" have to be handled like pencils, paper clips and notepads, which dont get from their manufacturers to your desk without passing through some distributors warehouse?
To those who have never questioned the underlying premises of data warehousing, the answer seems obvious. Its that data sets with differing structures need to be given a single, common structure in order to be combined to create (for example) master records of individual customers that can used by all of an insurance companys business functions.
Yet there are several problems with this answer, only one of which is that it isnt true.
But lets look at the other problems first.
To start with, once this erroneous assumption is accepted as correct, insurance companies facing the usual data storage challenges are condemned to settle for one inadequate solution or another.
For example, one fundamental drawback to any system that, in order to be usable, requires data from various sources to be given a common structure is that the system must be so complex that changing any part of it also requires an updating of the whole. And this updating, of course, must be performed by teams of highly skilled and expensive programmers.
Although such a problem is unavoidable in a data warehouse, it has come to be considered an acceptable limitation on a companys ability to use its data storage system, despite a modern firms need to respond rapidly and effectively to changing business needs.
Also particularly harmful is the way a data warehouse system mandates the homogenization of customer data originally created by different functions to serve their own particular needs.
Underwriting, marketing and customer service, for example, all gather and use very specific kinds of customer information, based on whats important for each of them to know. When such customer data has to be standardized across the company, the internal Balkanization from which many insurance companies suffer can lead to turf wars over how the conversion should be done.
Still another drawback to this standardization process is that when customer data from various functions is gathered into a central repository and "scrubbed" to make it available to anyone in the company, it inevitably loses much of its richness and particularity.
Yet in a business world where customer knowledge has become a key (if not the key) source of competitive advantage, this very richness is what gives such data its true value.
This kind of problem points to a broader limitation built into the very concept of a data warehouse. After all, you dont really want to warehouse your customer data, what you really want to do is to use it. And when you have new data to add to the system, you want everybody in your company who needs it to be able to work with it right away, without having to wait until you remodel your entire "warehouse."
One solution now being touted as an alternative to traditional data warehousing is called "unified data view" software. Yet even here, data must be taken out of its source databases and standardized, a la conventional data warehousing.
Whats more, it simply isnt true that, in order to pull together and work with data from disparate sources, one has to give it a common structure. Newer systems poised to replace the data warehouse concept can deliver to insurance companies systems the ability to:
Take data from any source–without cleaning, standardizing, or creating rules–and bring it onto one screen in the way that individual users wish to see it, not in the way that an information systems specialist says they should see it.
Allow users to work with data from disparate sources, with point-and-click workflow features cutting processing time significantly (for example, for policy applications and changes).
Attach a new database or administration system to a current data storage system without changing any old or new data–and to do so in a matter of hours.
All this means a revolution not just in the day-to-day operation of a data storage system, but also in its maintenance over time and through key changes in the business or the environment–for example, greatly reduced information system budgets and I/S-related overhead. (Even non-techies–that is, end users–can make changes to such simplified systems.)
As insurance companies continue to feel competitive pressure from the Internet revolution, financial services deregulation, industry consolidation and other forces now transforming the environment, they can no longer afford expensive, change-resistant, soon-to-be obsolete data storage systems that create enormous operational inefficiencies while hindering the pursuit of strategic objectives.
Companies need to start exploring alternatives to data warehousing that can enable them to streamline operations, dramatically cut costs and use their customer data to create competitive advantage.
Patrick Kassebaum is director of the Insurance Division of Mesa Corp. (www.mesacorp.com), a maker of enterprise interactive document-asset-management systems located in Lincoln, Neb.
Reproduced from National Underwriter Property & Casualty/Risk & Benefits Management Edition, September 21, 2001. Copyright 2001 by The National Underwriter Company in the serial publication. All rights reserved.Copyright in this article as an independent work may be held by the author.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.