Data Quality Assurance

No doubt we all desire quality data to work with, particularly when making strategic decisions. But how to go about assuring the data is indeed fit for consumption may be a non trivial pursuit in itself.

First, as with most corporate initiatives is the all important budget consideration. Assuming there is a rather compelling business case for it, i.e. we have executive buy-in, the question of who's going to do it needs to be answered - can this be accomplished internally or must we also rely on outside consultants with their broader experience base?

Next comes the daunting task of establishing scope - should we cover all corporate data or only pieces deemed mission critical? Even when starting small makes sense, there remains the decision of where to address data quality - downstream in target applications, within upstream data sources themselves or in a mix of the two?

Despite the challenges there is an approach to improving data quality that is both practical and frugal, viz., risk-based resolution. This prudent approach recognizes that some data quality problems are more probable than others and deploy limited resources to solve high risk ones instead of trying to mitigate each and every data issue.

Such a risk-based approach works because it focuses effort on where it makes the biggest difference - is there anything more practical than confronting problems likely to cause the most significant losses? - and by doing so, achieve a cost-effectiveness that is more palatable to the finance folks - hey, IT can be frugal after all!