Print Page   |   Contact Us   |   Sign In   |   Join Now
News & Press: SCIP eBulletin

Best Practices and Pitfalls in Data Governance and Stewardship

Thursday, June 14, 2018  
Share |

By Henrik Kjaer

Client Partner

SparkBeyond

 

Today, most organizations acknowledge the potential of data as a basis for achieving sustainable competitive advantage; some may even have the stated objective to become “data driven” which usually implies utilizing “Big Data” with its associated technologies and accelerated development methodologies to inform decisions about new products and services, customer experience, marketing, risk management and regulatory compliance.   

 

Deriving value from data requires effective Data Governance, i.e. the exercise of authority over the management of the data[1] and Stewardship i.e. formal accountability for the data.  

 

In this article I will discuss some ‘must-do’ tactics as well as pitfalls to steer clear of for IT and business leaders with a desire or need for Data Governance and Stewardship in their organization.  

 

Best Practices

 

• Relate Data Governance (DG) to Business Priorities and ObjectivesPromote data as the enabler in terms business can understand. DG teams must engage with business leaders to quantify the dollar value implications of the current state, the risks and impact of ‘Do Nothing’ to create awareness, and ultimately obtain a mandate to proceed. Derive DG requirements from approved business initiatives. Publicize wins aggressively to build momentum and expand the program

 

 Executive Sponsorship. The CEO of a major insurer I was interviewing at the inception of a DG program stated: “Our data household is chaotic.” I immediately knew we had traction! Whilst it is certainly possible to achieve pockets of success, without it implementations at the enterprise level require the active support of a committed (word and deed) Sponsoring Group comprised of key decision makers from the business with authority and influence to resolve cultural and people issues, secure funding and make resources available.

 

 Let the Business determine the data priorities and drive the approach. DGs role is as facilitator. Don’t resort to data profiling or your Enterprise Data Model to set priorities. It doesn’t tell you what your important data is. A carefully designed maturity assessment as a lead in can be a useful approach to identify gaps and gain agreement on priorities for the desired timeframe.   

 

• Suitable Data Owners and Stewards are senior members of the organization; they must be visible, respected and act as change agents. They reside in the business close to the point of data capture or consumption, not the IT department. IT stewards act in a supporting role, managing the IT systems and applications required for DG.   

 

 Formalize accountability for the data. Incorporate data responsibilities in job descriptions, performance reviews and incentive programs. Yes, this means getting your HR dept involved when you appoint Data Stewards. Align Data Stewardship with prioritized data subject areas, not source systems. 

 

Pitfalls/Obstacles

 

It’s not uncommon for IT leaders to claim ownership of corporate data improvement initiatives. Resist the urge; in my experience IT-led implementations are the least likely to succeed, although this may well change with the emergence of the next generation IT leaders.[2]

 

Don’t be overambitious. A well-designed DG program will identify ‘Quick Wins’ and deliver tangible business value early in the lifecycle but a greenfield organization without a history or culture of managing data as an asset requires a big (change management) effort that just can’t be achieved in a couple of months. It’s a learning curve; plan for the long term but effect DG one data subject area at a time to get hands-on experience and a better understanding of the potential benefits and risks. 

 

Failure to consider DG interdependencies. Data Science and Advanced Analytics are advanced practices built upon foundational DG building blocks such as Data Integration, Data Modeling, Meta Data (e.g. standard business definitions and metrics, data lineage) and most importantly, robust Data Quality Management. Bad data continue to hamper implementations, diluting the business value of Data Science and Machine Learning initiatives.[3] Maturity develops with time. You can significantly increase the likelihood of success by building out DG capabilities in a sequenced order, reflecting the dependencies specific to your organization’s needs.

 

Successful DG goes beyond tools and technology and embraces people and process change first. Successful implementations at the enterprise level require a programmatic approach to shift the perception of data and embed the planned benefits of DG during transition to the target state. 

 

BIO

 

Henrik is a consulting services executive with 20 years’ experience with the delivery of enterprise Data Integration, Business Intelligence and Advance Analytics solutions and services. He has worked as a strategic CIO/CDO advisor building data strategies, technology roadmaps and transformational programs for clients in telco, finance and manufacturing. Providing the leadership to make his clients business data-driven is his passion. 

 


[1] DAMA – DMBOK, 2nd edition 2017

[2] Pedja Arandjelovic, Libby Bulin and Naufal Khan, Why CIOs should be business-strategy partners, McKinsey & Company, February 2015 

[3] Thomas C. Redman, If Your Data Is Bad, Your Machine Learning Tools Are Useless, Harvard Business Review, 2 April 2018


Featured Members
Tracy A. Berry, National MI2018 Distinguished Member Award Recipient
Stephanie Hughes, Super-h, LLC2018 Innovations in Integrated Intelligence Award Recipient

Membership Software Powered by YourMembership  ::  Legal