As we all settle into a new year, it seems the demand for Data Scientists has not abated. The MBN team are still receiving plenty of new requests, and our partnership to place Data Science MSc students continues.
Demand also continues to outstrip supply. Despite the growing number of Data Science courses provided by UK universities, many leaders tell me they struggle to attract sufficient talent. For that reason, even as the leader of a successful and growing recruitment agency, I encourage others to think about alternatives.
Think about the balance in your team, not just one role
We’ve written before about the need to consider other data specialisms, not just Data Scientists. Consider the balance of your team. Are the skills gaps more in data access & manipulation (suggesting a Data Engineer)? Are the improvements you seek in presentation of the results of your analysis (perhaps a Data Artist)?
Chatting with some analytics leaders over coffee recently, brought home to me that many still need to hire data scientists. A number of them are in the process of expanding successful data science pilots. Moving from ‘chewing gum and string’ solutions, to a large number of potential business applications, needs more people. Plus, those additions need to be versatile, able to code Machine Learning, or other statistical solutions as required.
Could you have suitable candidates already working for you?
So, is there another way? It seems to require higher and higher salaries (or other benefits), to attract the in-demand talent pool, of trained data scientists. Instead, could you grow your own?
At first, that might sound overly ambitious. I was initially skeptical, when a dinner guest suggested the idea. But, they rightly pointed out the growing volume of resources, now available, to help people learn Data Science.
Universities are becoming more flexible in their delivery of content, as they compete with the success of free Massive Open Online Courses (MOOCs). Add to that the huge number of blogs sharing advice & resources, together with active local communities (including Scottish Data Science & Technology meetups).
Building on those sources, for understanding the principles and theories of Data Science, rich resources now exist for learning the most popular programming languages. From the success of EARL for the R language, to EuroPython for the growing dominance of Python (as the tool of choice). These events together with a host of blogs, hubs & online training, mean picking up such skills may now be possible.
Where internally should you look, for potential candidates?
If it is now viable, to empower someone inside your business, to pick up the knowledge and skills needed as a Data Scientist, where are they? Of all your potential employees, who should you consider?
Well, you will know your business and your people better than me. But, from our experience of screening and placing, thousands of candidates over the years, let me suggest a few possible pointers.
First, consider roles that may already have some of the skills you need from a data scientist. These ‘in demand’ individuals, are at their simplest, a fusion of data programmers + statistical analysts + business problem solvers. Married with a scientific approach, plus a curiosity to keep learning new methods and tools.
Such a combination suggests you might find potential candidates within:
- IT (esp. application developers or R&D teams)
- Customer Insight or Marketing Analytics teams
- Finance, Pricing or other technical analytics teams
- Process improvement or other technical operations teams
- Data quality management or other data focused teams
As ever, there is a lot of truth in the old maxim, hire for attitude, train for skill. So, I’d recommend not just focusing on candidates with the fewest knowledge gaps. Which individuals display the passion, curiosity & self-motivation to complete such a self-taught route? Have any already taught themselves to code in their own time?
Take a wider view of recruiting externally
Reflecting on that actually brings me full circle, back to the option of you recruiting externally. Your business is almost certainly already busy. It may be very challenging to release individuals from their existing roles. Plus, surprisingly, not everyone wants to become a data scientist (even if it is the sexiest role)!
So, how can our approach to potential internal candidates, help you take a different approach when looking externally? I’d suggest, by focusing more on what you require from a role, and the potential for imperfect candidates to learn what they lack.
The key may be in understanding what is essential, to your data scientist role, and which skills can be learnt more easily. Online resources, especially, are a very effective way of learning skills like programming. Mindsets, theories and an aptitude for creative problem solving, appear much tougher to learn through books and websites.
Reflecting on past successes, with placing candidates who ‘grew into’ the roles required by recruiting managers, I can see a few patterns. The first is the priority of attitude. You may need a data scientist to be an innovative & self-motivated problem solver, with the confidence to challenge accepted wisdom & ways of working. If so, finding evidence of such attitudes, or generic competencies, should be a priority.
Based on those we have seen go on to succeed in their roles, I’d suggest that the kind of skills that can be learned include:
- Programming in Data Science languages (R, Python, Julia etc)
- Using a relevant set of specialist libraries for those languages
- Potential algorithms to test for different business problems
- Different potential data visualizations & tools for producing them
- Managing and accessing data using different platforms/technologies
Investing in new candidates, a worthwhile investment
My final advice would be to continue to invest in developing new candidates. I sympathize that if it has taken considerable time and money to recruit a highly skilled individual, a business is hungry for results. My only caution is that people with these skills continue to be in demand. Those who choose to stay, often tell us it is because they are growing, with an employer who gives them time to learn.
So, whether or not you can hire the ‘ideal data scientist’ you are seeking, it might be worth considering how you can empower them to keep learning. This discipline appears to get more complex by the day, with new methods and tools competing for their attention.
The key to filling your data science roles, and keeping your data scientists, might be the same. Recognize the value of training. From softer skills to new tools, employers who provide time and resources for their employees to learn – they may well be the winners in the Data Science talent wars.