You may recall the 2012 article from the Harvard Business Review when it declared that being a data scientist was the “sexiest job of the 21st century.” In the time between then and now, more and more companies have taken on data science initiatives, data science degree programs have popped up across the country, and “data scientist” has appeared in LinkedIn’s Emerging Jobs report for three consecutive years.
Despite this promising outlook, you’ve probably heard that there’s a data scientist shortage. Companies can’t seem to find enough people with the right qualifications to fill roles. At the same time there are many applicants who believe that they have the skills and experience to be data scientists, but struggle to even get an interview.
Why is there such a disconnect? We think there is a misunderstanding about who is qualified to be a data scientist and the skills that they need on day one of the job.
At American Business Data Science (ABDS), our goal is to get more well-trained data scientists into the workforce. Our strategy is based on three pillars. We help employers:
- Redefine what it means to be a data scientist.
- Rethink how they recruit data scientists.
- Rethink how they train data scientists.
Data scientist, redefined
There’s no consensus on the definition of a data scientist, but web searches lead to descriptions like:
An individual with a deep understanding of statistical modeling/machine learning, computer coding, and industry/domain knowledge.
It’s time to throw out definitions that focus on specific tools and skills and think about what a data scientist actually does. Yes, they use statistical modeling and coding to help them reach their conclusions, but that’s not where the work begins or ends. First and foremost, data scientists solve problems.
First, they use critical thinking to find the right questions to ask. Next, they use their industry knowledge to formulate the right approach for an experiment. And when they’ve found a solution, they communicate it to non-technical audiences in an accessible way.
The standard definition of a data scientist implies that very few people could do this job, but that’s not true. Yes, people with technical backgrounds (data engineers, data modelers, and data analysts) are primed to be data scientists, but so are product managers, operations managers, and customer service managers. Both groups solve problems in their jobs everyday and already have the foundational skills necessary to succeed as data scientists.
Building the data science recruiting pipeline
When you change the frame of what a data scientist does, it expands your ability to recruit. Let’s look at all the possible avenues:
Cultivate internal data experts
You already have employees with some data science aptitude. Who are they? They are the ones spearheading improvement projects within their departments. They’re the Excel power users who depend on the program for most of their analyses. The only thing that they need is an introduction to coding and higher-level math.
On the flip side, there may be software engineers who are interested in business strategy. These employees can be taught how to apply their skills to tasks like business forecasting and planning, and learn how to better communicate with less technical audiences.
Become more involved in higher education
You may be wary of hiring students or recent graduates because they lack experience in your industry or because you think what they’re learning isn’t applicable to the workplace. Colleges are the perfect places to cultivate talent that meet your needs.
To ensure that students are graduating with the right skills, you can work with departments to develop more relevant curricula and offer guest lectures. In addition, you can give students experience through internships. When you are ready to hire for full-time roles, you have a list of people already vetted for their relevant skills.
Take a more holistic approach to traditional (external) recruiting
Talent sourcers should be on particular lookout for applicants who solve problems. This work can be aided by expanding your job descriptions, so that they attract people who see themselves as thinkers and troubleshooters rather than mathematicians or software engineers.
Requiring data science portfolios from applicants will give you insight into how they think. These portfolios should not be code repositories, but descriptions of business problems that they’ve approached along with results, visualizations, and presentations.
You should evaluate external applicants in the same way that you do high-potential internal ones. Look for their problem-solving skills.
Embracing data science training
Your new data scientist may not be “perfect” on day one, but that’s not a problem if you invest in training. This doesn’t mean you need to develop your own learning and training department. You can work with a vendor that has relevant courses that will meet your company’s needs.
That’s our specialty at ABDS. We know that the path into data science requires a holistic understanding of business, technology, and communication. Our courses are designed to increase a student’s skill in the areas where they need it most.
- Fundamentals of data science
- Business forecasting and planning
- Training for employees with technical backgrounds
- Training for employees with non-technical backgrounds
Our approach leads to a data science talent pool full of individuals with the skills, desire, and curiosity to help businesses solve the toughest problems.
Growing the Data Science Talent Pool was presented to the ANA, Association of National Advertisers, Data & Analytics Committee by Gildas Bah in New York City on February 6, 2020