Instructions to students: All questions are compulsory.
Dictionaries are not permitted.
Read this case study carefully before attempting Part A . You should allow 15 minutes for this.
The change problem at an insurance company (Company X)
Company X was an insurance company with two main departments. They had a call center to deal with insurance claims and a data mining department where terabytes of old insurance data were mined and the information used to help price insurance risks.They had a print room that would co-ordinate the printing, folding, addressing and distribution of letters. They had central databases, where all new data had to be stored. There were standardized protocols for the call center systems to interrogate insurance systems. There were executive information systems that consolidated reporting for senior managers.Despite all of this, some information was not easily visible.
The business rules for individual insurance systems were embedded in the software and could not be easily accessed by everyone. There were often change requests for new additions to internal systems. A typical change request would be, ‘we need a new system for this type of insurance because we need to be able to do function x that we can’t do at the moment’.In the business environment, Insurance Company Y were publishing what their target levels of market penetration were, but broken down by type of claim such as “car insurance claim – broken car wing”. Company X did not collect and publish this information and suddenly felt lacking in information to inform planning.As with any large organization, the definition of data items such as “customer” took on particular local meanings. This is not important until a local system needs to be integrated with a centralized system such as the central databases or by the post room.
The project to achieve a corporate data model did not make much progress.Because of the high levels of dependency between systems, there was a very formal change procedure applying to any change on the corporate IT network. Changes had to be submitted for approval and every Friday afternoon, when a group of middle managers met to agree changes. They averaged 300 changes per meeting. Informal estimates put the volume of changes that did not get submitted, but which were made anyway, at 1000 changes per week.