29Apr 08
Burning down the warehouse
Posted by Angus Kidman @ 15:38 2 comments
Getting executive sponsorship for any kind of data clean-up project isn't easy. If careful reasoning, detailed budget plans and a touch of blackmail don't work, then there may be a simpler solution: arson.
Want to know more?
For all the latest news, analysis and opinion on storage, click here
Before the man comes to bust my ass, I should point out that while that's not a serious suggestion, major disasters can often provide the incentive to get your internal data into shape.
Tony Fisher, CEO of DataFlux, gives the example of one company, a chemical manufacturer, which became a DataFlux customer after one of its manufacturing plants burnt down.
When the CEO realised that none of the internal divisions could provide an accurate list of customers who would be affected by this, thereby scuppering his plan to send an apology letter, a data cleansing project kicked off quick smart. (Sadly, I don't know whether the insurance would have covered this.)
Assuming you don't have a handy box of matches, then there are other possibilities. Fisher, who's visiting Australia this week to speak about master data management (well, somebody has to), is singing from a very familiar hymn book with his main suggestion: "Always look for the business case first. Once you've got a business case, getting executive sponsorship is much easier."
What kind of business case can you make for the inevitably tedious task of cataloguing all the data in a business, ensuring its accuracy, and building business rules that ensure it remains consistent?
Fisher suggests dividing such justifications into three central areas: revenue raising, reducing costs or risk management. Conveniently, we can label these collectively as the three Rs of data clean-up, though that may just confuse you in the end.
A little linguistic one-upmanship may also be useful. According to Fisher, while boards may find themselves heading quickly to Snoozeville if you mention "data quality", a subtle shift of phrasing to "data governance" apparently produces the corporate equivalent of a good shot of coffee. It's definitely a possibility worth considering.
But for all that, don't be surprised if it takes a lot of persuasion. Fisher is in the business of selling people on the concept of data quality, and even he estimates that awareness in business of the problem is only around 20 percent. It's enough to make you take up smoking, if only to have easy access to a lighter.
3%
2%
Yes, governance is the key. Shifting the onus of responsibility for data quality from IT to business management where it belongs. In the chemical company example who can be blamed for the lack of clean, integrated customer data if this was never put forward as a requirement?
Justifying data quality should be simple. Its just something that has to be done to meet business strategy. In the above example if a disaster recovery stategy was in place that stated that affected customers would be notified then the ability to identifiy customer information should have been considered. The implementation of strategies such as that then becomes the driver of data quality.