1. Case Study : Public Electricity Supplier

As the utilities markets became competitive their legacy systems underwent radical changes, effectively de-merging systems to be run by other, smaller business units under the new regime. From then on, the systems would communicate with each other using data flows – interfaces – conforming to a nationally-defined protocol.

It soon became clear that there was a scope for discrepancies when carrying out the demerging operation. This was because the extract mechanisms for producing the initial flows could be run at different times for different purposes, but the day-to-day legacy reading and billing cycle had to continue.

Datamotive’s task was to concentrate on business data quality by comparing the contents of the various subsystems on a “before and after” basis, and

analyse data across the component subsystems and interfaces to detect discrepancies

determine how significant was each difference in business terms

investigate the significant differences and associate them with their root sources and hence causes

offer plans and strategies to resynchronise the data

This was archived by designing auditing software to run in several applications within the operational cycle, and collect the results in a database with a common format. A reporting phase was run to show the discrepancies by age, and these were presented to technical and user staff alike to consult on perceived data quality. Following these meetings, remedies for those discrepancies which warranted correction were specified, coded, and implemented.

A feature of this operation was the practical recognition that the decades-old million-plus accounts on the legacy system contain discrepancies, but that in business terms, these were often immaterial. For example, dates “0” or “-1” might mean “a date in the past – the year dot” or “from now until further notice”, but provided that readings were being correctly recorded and bills produced for the correct amounts then it was probably not cost-effective to commit coding and heavy test and hardware resources to alter the situation.