Several sessions this afternoon on data quality and governance. Rather than blogging these separately, here are some thoughts:
- Great illustration of data quality problem having a business impact – bad data led a Telco to prepare a large CapEx project to add bandwidth capacity but a physical inspection showed plenty of actual capacity. Bad data had led to an unnecessary plan.
- An example given was that 20% of customers generate 80% of revenue so a loss of 1% of these good customers through bad data might make a real difference. Of course, if you don’t differentiate how you treat customers then it may not matter if you are wrong about who the profitable 20% are! Good quality data only becomes valuable if it is being used to make a difference in business terms.
- Funding must be linked to strategic imperatives – show that better data is either necessary for an initiative or that it would boost the results of those initiatives. Data quality is not likely to be funded directly.
- A lack of trust in information undermines data-driven decision making…
Several sessions this afternoon on data quality and governance. Rather than blogging these separately, here are some thoughts:
- Great illustration of data quality problem having a business impact – bad data led a Telco to prepare a large CapEx project to add bandwidth capacity but a physical inspection showed plenty of actual capacity. Bad data had led to an unnecessary plan.
- An example given was that 20% of customers generate 80% of revenue so a loss of 1% of these good customers through bad data might make a real difference. Of course, if you don’t differentiate how you treat customers then it may not matter if you are wrong about who the profitable 20% are! Good quality data only becomes valuable if it is being used to make a difference in business terms.
- Funding must be linked to strategic imperatives – show that better data is either necessary for an initiative or that it would boost the results of those initiatives. Data quality is not likely to be funded directly.
- A lack of trust in information undermines data-driven decision making. If people don’t trust it’s accuracy then they won’t use it, or analytics based on it, to drive their decisions.
- Suitable for purpose – which questions do you want answered, which decisions are you going to make, with this data? Use that to drive quality plans
- Analytics require data governance just as they require a level of data quality – it is hard to complete using analytics without governing the underlying data
- Regulatory requirements drive data quality, data governance – must be able to meet certain standards
- Drive the scope of your data governance program based on your data maturity, organizational structure/autonomy, external/internal influences/regulations, and the degree of executive support and drive – don’t get ahead of yourself
- Business must own and drive data quality and data governance – IT must act as a custodian of the data and nothing else. This, of course,is true of rules and decisioning too.
- Measurement, measurement, measurement – measure quality, measure governance, use your BI and performance management infrastructure to monitor these initiatives just like you would any other business initiative.
- Don’t forget to modify individual objectives and measures to reflect your data initiatives
I heard lots of talk today, in sessions and out, about how hard it is to get business owners to value data quality. My view is that this is inevitable and that the solution is to tie data quality problems to business value. And, of course, if you can’t tie a data quality problem to any business value then you should question whether it is really a problem…