Data Governance – Deliver Fast and Deliver Often

by Nigel Turner, Principal Consultant, EMEA

As an experienced data governance specialist, I am often asked why many data governance initiatives fail.  There are many reasons for failure but having helped many organisations refresh and revitalise governance programmes that are seen as not delivering their promised benefits, a common problem often rears its head.  This problem is that in many organisations that have gone down the data governance path there is a widely held view that a key reason for failure is that data governance takes too long to demonstrate any real benefits to the organisation.  As a result the business loses interest, the programme loses momentum, and the governance team’s funding and resources become under threat. 

Is this problem inevitable in data governance?  The simple answer is no.  It’s not an issue with the data governance discipline as such, but it is all about how it is implemented.   If badly planned, implementing data governance can indeed be a long and drawn out exercise.  There are many reasons for this, including:

• Data governance is first and foremost a business change management programme. In most organisations that are new to data governance the central concept that the business and not IT should be responsible and accountable for data can be a cultural shock.  As a result people can resist the change and the whole attempt can get bogged down in resistance and conflict.
• Second, setting up the organisation, processes and roles required to make data governance work can be time consuming. Newly appointed data owners and data stewards need to be identified, trained and supported to start driving up data improvement in the data areas or domains they are accountable for.  Doing this across a large business can take time and effort.
• Finally, many data governance roles are part time, with data owners and data stewards often having to balance their existing day jobs with the added responsibilities that data governance can place upon them. As a result governance focus can be pushed to the side and carried out only when role holders have the time and energy to pursue it.  This adds to delays and loss of impetus. 

So what are the right and wrong ways to implement data governance, given the above barriers?  The wrong way is to fail to recognise these inherent blockers and instead insist on trying to implement governance across the whole organisation from day one, akin to creating the Universe from the Big Bang.  Usually there will be a small central data governance team trying to drive this out and very quickly their time will be spread out too thinly across many different parts of the business. Inevitably this leads to the central team providing poor support to newly identified data owners and stewards, and so everything can slowly grind to a state where any progress is at best variable and at worst non-existent.   Already governance is failing and seen to be failing.

So what’s the right way to implement data governance?  Instead of the Big Bang approach, governance is best introduced through identifying and delivering pilots and proofs of concept.  To do this it’s vital to:

• Gain an initial view on the data management problems that are really hurting the organisation, whether these are centred on data quality, master data, business intelligence etc.
• Select one or two of these problems and analyse them further by involving business and IT subject matter experts. Hold workshops to define the issues and their impact in business terms, derive potential solutions (business change and / or IT change), identify the business benefits of data improvement and agree a potential plan of action to deliver them.
• Create a project team to implement the changes. As part of this team identify the key business stakeholders and ensure that they become the de facto data owners and data stewards of the data domain which the problem is focused upon.   This team should be supported by the central data governance team. 
• Deliver the improvements and revisit the projected benefits to ensure they have been delivered.
• Create a permanent data governance function focused on the data owners and stewards to ensure the benefits are sustained and further potential data improvements are tackled.
• Importantly, make sure that the whole project is written up as a use case and can be used to sell the benefits of a data governance led approach more widely across other parts of the organisation. Seeing a real success story is a far better way of selling data governance to others who may be resistant or hostile than theoretical benefits.
• Finally, repeat the above in other data problem areas. And so on.

The advantages of this approach over the Big Bang are many and include:

• Data governance is implemented first in those areas where data problems are causing the most pain. Presenting and publicising the use case to senior executives and others will help to convince them to throw their weight behind the wider governance initiative.
• Delivering real business benefits early helps to convince the doubters that governance has value to the organisation.
• The central data governance team is able to support the pilot projects effectively, and so guide the project teams to successful outcomes.
• The key governance concepts of data ownership and stewardship are being introduced gradually and this helps to embed their importance within the organisation and help to manage the cultural shift needed to make data governance a business as usual activity.

Overall, this approach is strongly recommended to anyone about to start to introduce data governance or is struggling with an existing programme.  When getting data governance right, delivering fast and delivering often is the key to success and adoption.   

Note:  Nigel is an instructor for our Global Data Academy courses, specializing in Data Governance and Data Quality.