I started learning about data quality management back in 1992. Back then there were no conferences, limited publications and if you received an email via the internet the excitement lasted for hours.
Fast forward to today. We are practically swamped with data quality knowledge outlets. Sites like the Data Roundtable, OCDQ Blog and scores of other data quality bloggers provide practical ideas and techniques on an almost hourly basis.
We never lack for ideas and methods for implementing data quality management, and of course this is hugely beneficial for professionals looking to mature data quality in their organisation.
However, with all this knowledge comes a warning. Data quality management can only succeed when behaviours are changed, but to change a person's behaviour requires the formation of new habits. This is where many projects will ultimately fail.
Have you ever started the New Year with a promise to change your ways and introduce new habits? Perhaps the guilt of festive excesses drove you to join a gym or undertake some other new health regime. How was that health drive looking in March? How about September?
The problem of habit formation is exacerbated when we attempt to change multiple habits. Perhaps we want to combine a regular running regime with learning new skills. The result is often failure.
To change one habit takes a considerable effort. but we often forget this when starting out with data quality. Armed with our new data quality library of techniques, we get to work building new information chains, implementing new technology, improving data models, creating ownership frameworks, populating enterprise data dictionaries and documenting endless data quality rules.
All of these techniques are often critical to data quality management success but ask yourself – what kind of effect will these changes have on your fellow knowledge workers? Will they embrace these new habits or are you asking too much too soon?
In my experience it pays to focus on one data quality habit at a time.
In my first data quality role the first habit we formed was documenting the information chain. Once we could all visualise how chaotic our information processing chain had become, it was far easier for the team to accept that our old habits had to change.
We then introduced new habits, one at a time. We changed the ownership of data, making people accountable for each stage of the information chain. Then we started to re-engineer the chain itself. Finally we implemented data quality controls at each junction point of the chain. All this took time, but the results were profound and could never have been achieved if they had been implemented in one operation.
I believe that moving too fast, too soon, even if you have an armoury of data quality tools and techniques at your disposal, is a recipe for disaster.
We are creatures of habit so focus on the most important habits first. Knit each one into the fabric of the workplace and then move on to the next. It will create a longer lasting pattern of change that is far more sustainable.
What do you think? Has your organisation had success with implementing multiple data quality habits in one hit? Welcome your views.