Human Capital
22
circumstances, the skills community would be able to measure individual and team value by multiple methods.
The environment inside most organizations is not ideal, however, for the establishment of a big data operation. Regardless of the level of challenge or the time required to implement the change required, our belief is that most organizations’ greatest advantage is awareness of their own limiting factors and their willingness to address these challenges to establish a deeper quantitative capability.
As one compares the ideal, unconstrained solution to the optimal, practical one, it is essential to think about the big data operation and the people who are members of it in a systematic manner. The human capital components of this community must work interdependently to get the most out of the decisions and investments that the organization makes in a data scientist skills community. This system of independent and interdependency components is shown below in Figure 1.
The Data Scientist Workforce Plan
Optimally, a workforce plan for data scientists will be requirements based. Establishing valid requirements without a data strategy (and without a Chief Data Officer) will be difficult, however. Without a data strategy the baseline workforce requirements to use in the construction of a workforce plan could be the current data science work being performed inside the sub components of the organization and (if it exists) at the corporate level. An inventory of functions currently being performed in each of these areas will enable all stakeholders to address what data scientist functions should be performed at the corporate level.
With an understanding of centrally controlled functions and decentralized independent work, organizations can determine the number of data scientists required to accomplish the current work. Although it would be better to build a projection of future work needs, taking an inventory of current functions provides an opportunity to establish the community around the current, known requirements.
With an understanding of centrally performed work (hub) and de-centrally performed work (spoke), an organization can document the force structure requirements by competency level and location. In one step, it could draft a workforce table which includes the organizational components each data scientist should be placed into. Ideally it would add data scientist positions to an existing corporate-level staffing document or by creating a new corporate-level staffing document in which to nest the data scientists and their teammates.
The collective effort to build the workforce plan to determine how many data scientist are required and where they reside is not as important as the sum of the human capital lifecycle aspects. Elements of the workforce plan should be evaluated on their accuracy as requirements and as to how they interact with and support the other components of the data scientist human capital strategy. A clear understanding and documentation of the competencies and the level of proficiency required to perform the work of a data scientist is essential to four other components of the human capital strategy, namely: develop, assign, lead, and measure. The most useful workforce plans include both what the organization needs and what it can resource at any specific time.
Determining the requirements and documenting them in detail is important to the success of the community and, in a small way, helps reshape the long-standing cultural preference for qualitative analysts. Success of the skills community rests on the ability to create a new “whole person” analyst culture where qualitative and quantitative contributions are valued. Without a solid base of competencies and levels of proficiency required inside those competencies, it will be difficult to establish credible business rules around evaluating
Figure 1: The PRIDALRM