Direct Taxonomy Reaps Many Indirect Benefits
Insurance Networking News, December 1, 2008
While many people may think taxonomies are mostly the topic of a boring 11th grade biology class lecture in which they dozed off, they're missing out on something that can bring great value to the business. Just ask The Hartford, which recently underwent a major initiative to modernize and unify their document repositories, but recognized that in order to maximize its benefit to the business, taxonomies needed to be incorporated into the plan.
According to Seth Earley, president of Earley & Associates Inc., which the Hartford, Conn.-based insurer contracted to assist with its plan, an enterprise taxonomy is a system for organizing and classifying information that can be applied to content management systems, document management systems, workflow systems, transaction processing and business process management applications, among others. The creation of a taxonomy is especially important when it comes to integrating information, as it provides a common language for describing information and consistent organizing principles.
Advertisement
"Taxonomy also is, to me, an understanding of information, a documentation of that understanding and then the application of that information," says Seth Maislin, taxonomist for the Concord, Mass. consultant. "Having a taxonomy means that you understand content, the context of how it's used, the relationships, the language; and it can be applied to documents and document sharing, search, the Web and processes."
Where taxonomy really comes into play for carriers is in the areas of findability, and making information more useful to individuals. "There's all kinds of improvements around Web site applications," Earley adds. "Insurance companies also can improve claims processing with the aid of taxonomy, as a claims processor can more quickly find answers, or a policy, procedure or guideline for customer service reps to process a claim when customers call. In this case, it improves the customer experience as well."
LOST IN TRANSLATION
The original plan for The Hartford was to move about 200 million documents into a new Documentum repository from EMC Corp., Hopkinton, Mass, but the insurer quickly realized after initial discussions that the transition was a prime opportunity to not just shift the existing metadata structure to a new place, but also to rethink how those documents were classified and organized.
"Everybody agreed the creation of an enterprisewide taxonomy would be the right thing to do, and we sold that to senior management," says Jeff Auker, AVP, eCommerce for The Hartford, Personal Lines, and the point person for the project. "They were sold on building a structure for the long run that showed a real commitment to doing things from a more shared-service, enterprise point of view."
After determining that taxonomy creation wasn't something that could be done in-house, The Hartford issued a quick, wide-ranging RFP that eventually led Auker to Earley & Associates.
Auker says he charged the consulting firm with two tasks: First, to set up a broad, universal taxonomic framework with about a dozen facets that would be applicable to any document The Hartford would use in its property/casualty business. The next was to assume control of the document migrations already in-flight, and apply that broad taxonomy to a particular set of documents-basically, do a deep-dive to refine the associated categories and metadata, which would help develop the search and access interfaces.
In order to construct the universal taxonomy, Earley & Associates conducted interviews with everyone from the executives who owned the processes (to define what is it they wanted to do with the documents) to the actual users in the field (to describe what is actually done with the documents). From there, the document bases were scanned to extract the metadata, and that information was used to create and manage the taxonomy.
HIERARCHICAL DISARRAY
One of the major challenges at the time, Maislin says, was in dealing with the documents the underwriters needed to create/receive in order to activate/maintain a policy.
The Hartford's original system used the terms "template" (meaning all outgoing documents) and "attachment" (meaning all incoming documents) to classify its underwriting documents, but of those 18 million documents, 14 million were just lumped in a folder labeled "other."
"Over time, people got away from using the existing classification system, and just threw everything into a big bucket, so no organization took place," Maislin explains. "When it came time to identify what something was, you'd have to talk to the underwriters and customer service, who had to review a huge number of documents on a policy in order find something relevant - they looked at docs by their dates, hoping they could guess, based when the document was written, if it was the right one. So to remedy this, we created a structure to map the documents that would make sense based on what the underwriters needed."
For more information on related topics, visit the following channels:
