Prediction – Data Science Skill

Prediction

Management roles in data science will remain in demand for years to come, but their leading contribution will shift from being a stop-gap in the data analytics pipeline (as seen today) to leadership for developing and managing resources across it. The practical implications for organizations looking to develop a new data science capability, or grow an existing one, are: use consultants to get started, develop existing employees, organize interdisciplinary teams, and build for the long-term.

Article

The job market continues to be hot for data science resources. Characterized by a unique blend of programming, statistical and critical thinking skills, data scientists are in demand because they work across traditional gaps between IT, business, and executive leadership. Finding a true data scientist is difficult, however, as many candidates have deep technical knowledge but lack the soft skills necessary for effectively collaborating with line of business owners. Accordingly, McKinsey & Company forecasts that the demand for data scientists will continue to grow, potentially reaching a level twice as great as supply by 2018.

As one might expect with such a forecast, the market reaction has shown a dramatic increase in educational offerings for technical skills requisite to the data science role. Unfortunately, such efforts are destined to fall short. TDWI research (Harper, 2015) reiterates that the top three qualities businesses want in their analytic roles are:

  • a knowledge of the business,
  • a familiarity with data,
  • and a critical thought process.

Understanding how to manage data, write programs, and run statistics is quickly becoming table stakes for all analytical roles. The real premium is for business minded individuals who can articulate insight from various forms of data.

The real question we must ask then is: should we train up statisticians and computer scientists to think like business owners or should we train up MBAs to work more like statisticians and computer scientists?

Amazon, a company with a data-driven culture, has chosen the latter. All of their business analysts are trained to use structured query language (SQL) for doing their analysis. Many also use statistical programming languages like SAS or R.

Google, a company founded on data, has traditionally run a flat organization with few managers and many technically deep resources. Their model implies a low ratio of managers to overall technical roles, thus suggesting, like Amazon, that it is more effective to train the MBA to work like a statistician than it is to train the statistician to think like a business owner.

These high tech examples provide long-term insight for any business who has an emerging data science need. Namely, it will be a competitive advantage for you to develop business talent as data scientists than it will be to put programmers into business advisory positions. Indeed, most companies have resources already familiar with the business and the data. What is missing is the appropriate development programs and the leadership needed to facilitate organizational change.

Therefore, it is my prediction that management roles in data science will remain in demand for years to come, but their leading contribution will shift from a stop-gap in the data analytics pipeline to leadership for developing and managing resources across it. The practical implications for organizations looking to develop a new data science capability, or grow an existing one, are:

1. Engage consultants upfront to build strategy and to execute on mission critical work.

2. Invest in the technical development of internal resources who possess both business knowledge and data familiarity.

3. Build interdisciplinary teams founded on a common programming language (e.g., organize a DBA, statistician, finance partner, and business partner – who all know SQL –  into a single team).

4. Think long-term. Companies not founded on data will have a persistent problem recruiting and retaining true data scientists. Invest in an adaptive, long-term strategy and develop new talent accordingly.

As is the case with most market phenomenon, what goes up, must come down. Data science is here to stay but the hype over data science skills will slowly give way to the underlying need for soft skills, entrepreneurialism and executive communication – the keys to effectively deploying analytical insights. Like the current state of higher education, where the average student has $15,000 in debt and few employment prospects upon graduation, so will it go for the run-of-the-mill technical resource aspiring to be a data scientist. Recruit soft skills, develop technical skills, and promote your employees who know the business and the data into data science management roles.