KDnuggets Home » News » 2016 » Aug » Opinions, Interviews » Is a Chief Data Officer Required for Analytics Success? ( 16:n30 )

Is a Chief Data Officer Required for Analytics Success?


In this insightful opinion piece, gain perspective on whether a Chief Data Officer is required for an organization's analytics success.



By Gahl Berkooz, Global Chief of Analytics, General Motors.

Header image

The Data Organization, headed by the CDO, typically operates in three dimensions:

  • Governance – identifying, developing, and promulgating data definitions, data standards, and data quality metrics. Once developed, a good CDO will create score cards that measure data quality by business function, and utilize enterprise data governance to hold the business functions to continuous improvement of data quality.
  • Data preparation for Analytics – the adage that 80% of the effort of an analytics project is getting the data ready is true. A CDO often serves the Analytics teams by loading, cleansing, transforming, and packaging data. This makes the Data Scientists/Mathematicians more efficient, and increases the throughput of the Analytics teams.
  • Solving business problems with data root causes – A CDO can take on the role of “owner of the data assets of the company.” As such, s/he will seek to improve and buttress the asset. Where there is business inefficiency as a result of data issues (a.k.a Data Waste), or an opportunity for business innovation through superior data, the CDO will work with the business functions to deliver the opportunity. This results in data standards and improvements to data at the source, the transactional systems used by the business function. Assuring data quality at the source in required to sustain the efficiency or innovation, and reduces the Data Waste analytics projects need to contend with. A good description of how it was done at Ford Motor Company is available HERE.

For Analytics to be successful in the long run, i.e. for Analytics solutions to be deployed in an embedded and ongoing decision support capacity, data disconnects need to addressed at the source. The Data Waste that required 80% of the effort in Analytics projects needs to be gone for good. Data must arrive in a “clean” state from the original transactional systems or sensors (in the case of a connected product.) Analytics that rely on the Data Organization continually, and manually, cleansing and transforming data does not scale: every new Analytics capability results in an incremental task for the Data organization, an additional “pancake” on the stack.

How can the CDO eliminate data disconnects at the source? The simplistic answer is through Governance. In theory the Data Organization will set standards and hold the business functions accountable to following them. This is easier said than done – data inside an established business function has evolved over time to optimize the operations of that function. Modifying the data to comply with new standards has a cost and a potential adverse impact on efficiency in the originating function. The argument of “do the right thing for the enterprise” can be invoked, but that will be followed with a discussion about the business value of the Analytics versus the cost and loss of efficiency. The way to avoid getting bogged down in such a conversation is to eliminate Data Waste in a way that is a win-win.

When the Data Organization solves business problems with data root causes, it has the opportunity to set information standards whose implementation is a net-gain to the business functions. This is the third dimension for the Data Organization, as mentioned above. How to build this capacity in the Data Organization is a conversation on its own (a description of how it was done at Ford Motor Company is available HERE.) Once the operational benefit from Data Waste elimination is delivered, the business functions get a return on the investment of cleaning up information at the source. This provides the business functions with the motivation required to clean up the data and adhere to standards. As described HERE, building a Data Organization that eliminates Data Waste in the business functions is challenging, yet rewarding to the company and the team.

So, the answer to the question is: depends. The make a sustainable impact on an Analytics, the CDO needs to create self-sustaining momentum of elimination of Data Waste in the company as a whole - not just focused on Analytics.

This post reflects my personal views. The views herein do not represent the views of my employer.

Bio: Gahl Berkooz, Ph.D. is a Global Chief of Analytics, Global Connected Consumer Experience (GM/GCCX) at General Motors. He is based in Detroit area, MI, USA.

Original. Reposted with permission.

Related:


Sign Up