The Accident Compensation Corporation (ACC), with the help of data analysis specialist SAS, has begun a project to raise the quality of its data, eliminating inconsistencies, inaccuracies and duplications, as well as ensuring consistent naming of data fields across the corporation’s various departments.
The effort sits within a larger strategy of improved data governance in ACC, which involves structural changes in the organisation. Data quality is often seen as an IT or systems issue when it is really an organisational issue, says information services manager Jane Hardy. Once a commitment has been made to improve data quality, “the business processes have to reinforce that”, so errors don’t creep back in and have to be weeded out on an ad hoc basis.
Data quality improvement is recognised as an obvious thing to do, but taking action throughout an organisation to inculcate a more systematic and mature view of the data lifecycle is not as easy as it might seem at first sight, says Hardy. ACC’s approach to data quality has been “going through a maturity process that a lot of organisations go through”, she acknowledges.
Ensuring consistency of data within ACC will assist the organisation to get a more complete view of its clients – for example, linking a client’s claims history with the record of their levy payments – as well as improving standards for data flowing between ACC and other government agencies, such as the Inland Revenue Department and Statistics New Zealand.
“We’re moving more towards an insurance model, which means linking the data in different ways,” says information strategy and planning manger Zeeman van der Merwe.
ACC is looking at opportunities for other kinds of cross-matching that better-managed data quality will facilitate, for example matching with police records of road accidents to assist in reducing the accident rate. Following a Law Commission analysis of data cross-matching between government agencies, it is likely to become easier to match data and glean valuable information.
The ACC effort will begin with ensuring consistency of customers’ addresses, ensuring they all conform to standard New Zealand Post format. This means the genuineness of the address can be more easily verified and it can be attached accurately to geospatial coordinates and Statistics NZ’s meshblocks.
ACC has been using SAS for various kinds of analysis for 25 years, but in this case it is using tools not from SAS proper but from data-quality specialist DataFlux, which SAS acquired in 2000, and which has remained a distinct unit within the company.
The selection was made after a restricted tender with three suppliers on the shortlist, says van de Merwe; he declines to name the other two shortlisted candidates. As well as being an acknowledged leader in the data quality market, the DataFlux tools proved “better integrated” among themselves and with related software than are competitive offerings, he says.