In a recent post the very knowledgeable Nicola Askham suggested that an enterprise ought to have the fourData Quality roles of Data Owner, Data Steward, Data Produces and Data Consumer. Though I agree with Nicola that these roles are valid early on, I also believe that, if they persist, then the enterprise is very unlikely ever to achieve true maturity in data and information quality.
My full reply is posted below.
I do agree that these roles are essential in an organisation where data and information quality is not very mature. However, if they persist, then that enterprise is highly likely to always have a primitive quality control, as opposed to a quality assurance, approach to data and information quality.
The role of Data Steward, in particular, is a real poisoned chalice. This person will seldom be senior in the organisation and is quite often expected to perform this thankless task, which is really all about cleaning up other people’s mess, often without being paid for it, while at the same time having absolutely no control over the people creating the mess.
A quality assurance approach to data and information quality is the only truly sustainable model. In this model it is the users of the data – those people who convert it to information by performing a Business Function – that own and are responsible for defining the format and structure of the data that will meet their needs. It is people performing the functions that create data that are responsible for creating it in a manner that ensures that it meets the needs of those who will use it, i.e. that it will be fit for purpose – the fundamental definition of quality.
Managing quality by data ‘sets’ is not very effective as the execution of a single Business Function can often validly result in the creation of new occurrences in several different ‘sets’.
Giving ‘ownership’ by sets can be highly problematic as one executive might want to own ‘Customer’, another to own ‘Supplier’ and a third to own ‘Employee’. It was such assignment of ownership that resulted in the catastrophic split of the master entity of Party, that required the whole discipline of MDM to be invented. A discipline required to undo fragmentation that should never have occurred in the first place.
Thank you for starting the discussion.
What are your thoughts on such roles?
If you would be curious about finding out how I can help you move quickly through the various stages of Data and Information Quality maturity in your enterprise, please contact me: Email: firstname.lastname@example.org Skype: johnowensnz Phone: +6421774785 (NZ time 8am to 8pm)