Coming up with a business case for a data mining strategy might be a tricky business but if it generated AU$118 million in additional revenue, then it would probably be something of a no-brainer.
AU$118 million is the amount of overdue taxes that the Australian Taxation Office claims to have collected by using data mining tactics to identify trends amongst taxpayers with outstanding returns.
The project, detailed in an article in today's AFR, assigns risk scores to different taxpayers based on past behaviour and uses that information to determine the most appropriate strategy for seeking payments.
Serial offenders might immediately be issued with a summons, while first-timers will receive the standard letter used on such occasions.
While the fiscal and regulatory outcome is a worthy outcome in anyone's books, replicating the experience in other businesses might not be so easy, and not solely because no other business has the right to demand payment from virtually every business in the country.
The ATO's data mining project had a particularly large scope, since it had access to databases from other organisations such as Centrelink and banks as well as its own information.
While that makes the initial task of filtering more daunting, it also offers a richer resource for tracking behaviour that isn't readily accessible by the private sector.
The ATO also has a simple and useful tool to help it deal with the perennial challenge of data quality: tax file numbers.
Being able to track individuals via a single assigned number makes it much easier to reconcile problems such as "Is this John Smith the same as that John Smith?"
So while the ATO stands as a good example of data mining's benefits, its experience is probably more relevant to government CIOs than their corporate colleagues.