Analytics
21
In Human Capital Management: Part One
Challenges
There are five major challenges or barriers that will inhibit the establishment of data scientist talent pools and their contribution to an organization’s mission.
•
•
•
•
•
Some organizations place a heavy reliance on qualitative work. Data scientists act in a supporting role with the efforts of the qualitative analyst. Some public firms tend not to reward or recognize collaborative contributions between two or more analysts. Without proper recognition, morale and productivity will remain low for data scientists.
Another issue that holds organizations back is access to data. Data is like oxygen to a data scientist. They must have large quantities and uninterrupted access to perform their work effectively.
The third factor complicating the establishment of a data scientist capability is singular management control of the capability and the people who provide the capability.
The fourth limiting factor is the lack of an accepted data strategy. Without a clear data strategy that can be used to establish the data scientist’s contribution by way of competencies and expected return on the investment, the best that can be achieved with this group are contributions that are not contested and applied in a general way to or appreciated at the highest levels of the organization.
The final factor that will limit a best-in-class solution for a data scientist community involves timing. If organizations wait too long, an already thin talent pool will be even more elusive. Moreover, the organization will fall behind its peers in leveraging analytics for better decision-making.
A Data Scientist Human Capital System
An ideal data scientist talent pool will require the design, development and implementation of a data scientist human capital system, which combines as eight contributing components of the talent management lifecycle. In this article, we use the PRIDALRM talent management lifecycle to represent Plan, Recruit, Inspire, Develop, Assign, Lead, Retain and Measure (the components of the talent life cycle are discussed in further detail through this article).
Our research suggests that change of the magnitude necessary to implement a big data capability takes two to three years at small or mid-sized organizations and longer at larger agencies and organizations. However, it is our belief that skillful leaders of change can implement and adapt our recommendations given adequate time and resources.
The “ideal” situation for the creation of a data scientist skills community is different from the “optimal” solution. Where ideal represents what can be done in the absence of real-world constraints, the optimal represents what can be achieved within the known constraints. An ideal solution might, for example, be designed around an accepted data strategy. It would reside inside a culture where qualitative and quantitative analysts were valued proportionately, based on processes and products produced in collaboration.
The ideal big data operation would be managed at the corporate level and be led by a senior executive (Chief Data Officer) with adequate cross-corporate authority. The community’s members, as a whole, would have the confidence of the leaders and the clearances necessary to access all of the data they need to do their work. This community would be resourced with a recruiting pipeline offering established visibility to potential talent as early as high school, and with college-aged candidates participating in an active, long-term talent acquisition process. It would include funding for technical training and education with a total compensation structure comparable to the largest public or private sector employers of data scientist talent. Retention would be driven by the commitment of the leadership to provide access to data, and to challenge the scientists to continuously find new and innovative ways to enhance the contributions of intelligence processes and products.
Ideal development of this talent would include diverse assignments and continual learning to maintain skills in data mining software. Talent development would include collaboration between the members of the data scientist skills community. This collaboration would allow data scientists the opportunity within “cleared to share” boundaries to champion their work and hear about other data scientists' work. Under ideal
By: Allen Zeman, Ray Horoho & Tom Myette
Center for Human Capital Innovation