With over 10 years experience working solely in the Data & Analytics sector our consultants are able to offer detailed insights into the industry.
Visit our Blogs & News portal or check out our recent posts below.
As such, the Data Scientist is the hottest and most sought-after job title in today's job market and competition has never been so fierce. Ultimately, competition is only going to get more and more intense for these talented individuals.
So how do you win the war for Data Science talent? Over the course of this talk, we're going to look at talent attraction and retention strategies, how to engage hard to reach talent, and how to create stringent and effective hiring processes that will differentiate your business from the competition.
Is there a shortage of talent?
The answer to this is both yes and no. Run a quick LinkedIn search for a Data Scientist in the Bay area and you'll see a massive 35,783 returned results and 2,604 live roles, however a 2012 government study stated that undergrad stem degrees would need to increase by 34% in order to meet predicted demand for these skills.
However at the same time, only 50% of individuals with a STEM degree are employed in a relevant field - this number is roughly 3x higher than other fields of study.
32% of Comp Sci degree holders are not employed in an STEM role say that they are working in unrelated fields due to a lack of relevant job.
The reality is that for the most part - many of you in this room today are speaking with the same candidates as one another and are looking for the exact same profile. At the same time, candidates are all applying to the same companies, with the same focus, again, for the same roles. The same companies are offering jobs to the same pool of candidates, who have 4 or 5 job offers in hand at any one time, meaning that 75-80% of job offers aren't being accepted.
A vicious cycle
This is leading to the cycle that I see in the market, every single day.
1) A large proportion of candidates are overlooked or excluded, giving the perception of a lack of opportunity
2) Employers are struggling to fill their roles due to losing candidates to competitors, feeding the idea of a lack of available talent.
What's the Fix?
The first step is looking beyond the standard candidate pool - if I were to ask all of you to think of your ideal data scientist, most would say, Masters or PhD in a quant field. Internship somewhere scrappy during the course of their degree. Publications, Patents and Research work. Few years of experience in a production capacity, Python, R, Strong ML background etc etc.
Maybe it's time to look beyond that - Organizations such as Galvanize and Metis have created intensive data science programs that are creating well rounded data scientists that most people will look beyond. Let's change the focus of the degree criteria - some of the best coders that I've seen, have been self-taught and have a genuine passion for developing their skills.
So how do we find these people?
There will always be a time and place for LinkedIn's searching features, but the candidate pool runs so much wider than that. Luckily, you're all at a conference, so I don't need to pitch you on the importance of networking, but there are so many ways to find this talent. Never underestimate your own people, the best people, know the best people. Before you do anything at all, brief your team on what you're looking for and open the floor to them for their referrals and recommendations.
Next, create partnerships. Work closely with schools, bootcamps and research facilities to get you access to data science talent quickly and efficiently. You'll also be able to get a fresh approach to those problems you've been looking at and may just stumble on some exciting solutions.
Sponsor a competition. So many Data Scientists that I work with are fiercely competitive and love the idea to showcase their skillset. You'll also get a natural interview process taking place, with the cream rising to the proverbial top.
Lastly, keep up to date with the market - follow market moves, funding rounds and news stories to look in to redundancies or news stories that may make an individual more open to a move.
Finally actively target companies that utilize similar tools, or work on transferrable problems. By no means am I saying headhunt your competitors staff, but why not look at areas that will utilize similar methodologies or algorithms, where someone will be able to come in and hit the ground running.
Engaging with this talent - The power of why
I'll never forget the best piece of advice that I had received when it came to recruiting Data Scientists. It came from Vin Vashishta (who is definitely someone you need to be following) - Focus on the "why".
Your message needs to stand out, and you need to capture the imagination of your potential hire, especially if you're not a traditional "halo" brand. Candidates in this space are mission driven, the what is nowhere near as important of the why.
For example, candidates don't care that you're utilizing computer vision - but they do care that you're using it to monitor and track the breathing patterns of infants in their sleep. It's going to be the why that starts the conversation, the what will come after.
Engaging with this talent - Growth & Development
Next you need to be upfront about growth and development. Not every organization is going to change the world, not every start-up will achieve unicorn status. That's ok. Not every scientist will become a CDO. Don't make promises that can't be kept and be clear about where and how you see this role and this person evolving.
Engaging with this talent - Do away with lengthy processes
I follow very closely the work of some exceptional people, and I love seeing Data Science applied to hiring processes. One such person who I follow is Emily Glassberg-Sands at Coursera. She wrote an article about how they had analyzed every area of their hiring processes and assessed where people were dropping out, and fixed those areas.
Let me tell you now, you don't need to have 5 screening calls, a take home test and an 8-hour interview day. In fact, simplifying this even further, you don't need a take home test. Nowadays most people have a portfolio of code that they'll be all too happy to share, so that you can see first-hand the work that they've done. If you want to understand how candidates approach and tackle a problem, run a whiteboard session, or a webex, where a candidate can feel like they're already working with you, tackling a problem in unison.
In any hiring process, you're getting interviewed as much as you're interviewing. Your process reflects who you are as a business, long and drawn out - means slow and clunky. There's an organization whom I know of, who are doing a huge hiring drive, with an average turnaround time of 8-10 weeks per hire. As a result, candidates are getting half way through a process, getting messaged by another organization and are off the market a week later.
Focus on what is a necessity, a stage focusing on technical capability, a stage focusing on role suitability and a stage focusing on cultural fit. All in all, three stages should be more than enough, and should take no longer than 10 days.
Set expectations clearly at the beginning. Hiring is a time-consuming process and losing time interviewing a candidate whose expectations are not aligned with your own, is a waste of time, purely and simply. Have the difficult conversations as early as possible will save you time further down the line. Put bluntly, if/when you work with a recruiter, the vast majority of people ask us to find a candidate's salary expectations, however most organizations when they recruit for themselves, they do not have that conversation until the very end of a process.
Engaging with this talent - Closing candidates
The most effective and efficient processes mean nothing if ultimately you can't get a candidate across the line. This is where your recruiter - either internal or external is going to really earn their money. Understanding the push and pull factors in a decision is key to a successful hire. A high base salary means nothing if that isn't the reason that a candidate is looking for a move.
The key here is that when you've identified your hire, strip everything back to the bare bones. Go through the role again, the company again, understand any concerns that they may have and set up conversations with decision makers again if necessary. Close candidates on numbers at which they feel happy, but that also mean value for money for you - DON'T LOWBALL!!! I see all too often companies that can go higher, go in with a lower offer to try and get a discount on a hire, every now and again, you'll get a positive resolution from this, but more often than not, you'll scare off a candidate who has the potential to feel undervalued, and therefore warn other people in their network about a negative experience. Make offers that are fair, for both parties and explain how you got to those numbers.
The closing process is the most in the whole recruitment cycle. A botched close will mean that you start back at square one and that all of the work that you've done starts back at zero.
Retaining this talent - Prepare for churn
Unfortunately there is no secret sauce here. It's going to happen. By the very nature of hiring scientists, you're looking for people who are naturally inquisitive with a thirst and passion for learning and development. More often than not, you'll form part of their development, as opposed to all of it.
The average Data Scientist is currently switching role roughly every 2 years. Your job is to lift that number as high above the average as you can. The key here is to understand who your Scientists are as people, what's important to them in their future, and help to meet as many of their goals as you can.
Retaining this talent - Invest in your people
The old saying rings true - "what if we invest in our people and they leave - but what if we don't and they stay". You need to give your scientists the freedom and the platform to be the vest version of themselves that they can be.
Lastly, don't wait too late to reward your top performers. A counter offer is always an offer too late. If someone is performing well, let them know that they're appreciated with that promotion or raise that they deserve. Don't wait for them to come to you with their 2 weeks notice. If you wait that long, you're too late.
Retaining this talent - People leave bosses
I don't believe this to be a fundamental truth. As I mentioned earlier, I firmly believe that the why is the most important thing in developing great scientists. As long as your mission is one that excites your people and that you're constantly following your north star, getting closer and closer. Your scientists will be driven by that same mission.
Ultimately your role as Managers, Directors, VP's and Execs is to nurture the talent within your ranks, create an environment where your people can thrive, and where they know that they'll continue to do so.
Visit our Blogs & News portal or check out our recent posts below.