Assert common definitions
Asserting a single definition for each of the various data types is simple in principle, but can be difficult in practice
One of the fundamental tenets of good data governance is that common data definitions are observed across the organisation. Sometimes, says DQM Group’s Gregory, “it’s as basic as defining what a customer is.”
But something so superficially simple has proved to be highly challenging for many organisations. Howard says he has “come across companies with literally tens of definitions of the word ‘profit’.”
Gwen Thomas, president of the Data Governance Institute, recalls how one multinational hotel and leisure company operated with multiple interpretations of the field ‘length of stay’ for several years. While the problem did not directly hit the firm’s bottom line, it did gravely impact the accuracy of internal reporting.
Organisations have recently begun to move from an entirely “siloed” approach to IT systems and data, Thomas explains, to a world where information is shared across departments despite differing interpretations of the data. That has brought all manner of internal inconsistencies to light. “Data that was completely perfect for one purpose may suddenly become inadequate without any changes being made to it,” she warns.
It is the job of the data council to devise and implement policies that ensure data adheres to standard definitions across the organisations. Indeed, Thomas argues that this is one of the council’s first priorities.
Automate where possible