Data governance must encompass management of the full life cycle of a data policy – its definition, approval, implementation and the means of ensuring its observance

- David Loshin, Data Policies and Data Governance

I was checking out my Google stats on Data Quality Pro recently and observed that "How to create a data quality policy" was one of the most popular searches people make. It's easy to understand why.

As you start to mature data quality, you move away from doing isolated cleanups and begin to think about how you can change the wider organisation. Promoting change, even at a departmental level, will require some form of agreed mandate. This need for a template invariably leads to a hunt on the web for how to create a data quality policy.

Policy templates are a useful starting point – but from a data governance perspective, there can be far more value in going down another route.

Let me explain with a true story from my distant past.

My first introduction to data governance came via an ISO 9001 initiative with my employer back in the early nineties.

I had been happily coding data quality scripts (I started out as a software engineer) when I was summoned to the boardroom to talk about policies. My heart sank as I was duly enrolled (under considerable duress) onto the ISO 9001 steering committee. What followed was weeks of tedious discussions about policies and procedures.

Because our company was in effect manufacturing data products (instead of ball bearings or meat products), we had a lot of challenges aligning the traditional manufacturing perspective of ISO 9001 to our data-driven information processing plant. My frustration and boredom grew until one morning I had what can only be described as a Data Quality Eureka Moment (DQEM if you will).

During a policy formation meeting (yes, it was as interesting as it sounds) I realised the source of my long working hours was related directly to poor information chain management. By following the ISO 9001 policy of ensuring that all materials must be traceable to their source, we started to implement controls that quickly highlighted which suppliers were sending poor quality data. We soon built firewalls with checks and balances, to stop the data getting into our core information chains and creating havoc downstream.

The effect was dramatic. Everyone adopted the new policy. Customer delivery lead times tumbled. Profits increased. I got my life back.

You might think the moral of this story is that information chain management is an important data quality technique. It is – but this article is about data governance, not data quality. So what's the takeaway? Did you spot it?

Think back to what happened earlier in the story. The ISO 9001 leader (head of enterprise governance in today's parlance) invited me to attend the policy meetings. To say I was "green" was an understatement. I knew zip about policies, mandates, processes, controls and governance. But there I sat, week after week, soaking up all the nuances of ISO 9001 and how we could comply. And that was the key to the success of the "Data Governance Program" the company undertook.

I helped set policies and directives. I devised the templates and boxes for users to complete. I wrote the instructions. I came up with the strategy for tracking data electronically.

But it wasn't just about me. I had to continually go back to the team and ask for their input so that the policies would make sense in our language, our way of working. That decision to take someone from the lowest ranks, and then empower them, played a pivotal role in the program's success.

So guess what happened when I finally went back to the office and asked all of my coworkers to agree to work with the official set of policies?

They all agreed.

And the reason why stems from the quote by David Loshin at the start of this article:

Data governance must encompass management of the full life cycle of a data policy – its definition, approval, implementation and the means of ensuring its observance.

We all too often get wrapped up in the observance aspect of data governance and forget that without the right involvement and representation of the workforce, all bets are off when it comes to implementing policies – they won't care. But get the people integrated into the process and they will empower others to take action.

Tip for the day: Create a data governance policy that ensures full participation from the workers. Don't create data governance policies from on high and expect their compliance.

It doesn't work.

Download a white paper about big data governance