Today, opportunities are aplenty in the data science space. This is good news for junior to mid-level data science professionals, but for leaders of data science teams who already find it hard to attract strong talent, it is getting increasingly difficult to retain the brightest in your team.
This article is written by Ong Su Yi - Manager, Data Science & Technology. She has 6 years of recruitment experience across the data science, technology and financial services sectors.
As a data science recruiter, I have heard so many ‘push and pull’ factors. Some data scientists move for money, some move for progression. Some move to work with more data, some move for more interesting use cases. Some are really difficult to move because they report into a manager they highly respect.
With intense competition in the market and lots of cash to burn (at selected start-ups), what are the traits of a good manager which could reduce the probability of his team members jumping ship? Having met Chief Data Scientists across the market and their team members, I thought it would be interesting to share my observations.
Traits of a Good Data Science Manager
Thinks like a Businessman
Data scientists love working on meaningful projects which impact businesses. Be it driving revenue, optimising healthcare resources or reducing man hours spent on reading documents. Most data scientists don’t just want to develop proof-of-concepts, they want to see their novel ideas adopted by the business and make real impact.
By being able to think like businessmen, data science managers are able to reassure business teams that the data scientists are not here to steal their rice bowls. Data teams are built to complement business teams and help them do their jobs better. Data science managers who talk like businessmen are more successful at getting stakeholder buy-in, securing resources (especially budget) and drive adoption of analytics products within the markets.
In today’s world where the customer is king, you don’t want to overpromise and under-deliver. Good data science managers understand the challenges of developing analytics solutions, and is able to manage the expectations of his business stakeholders. After all, you don’t want to overwork your data scientists and motivate them to click that ‘apply’ button on LinkedIn.
These leaders should understand the challenges of developing analytics solutions, and be able to manage the expectations of his business stakeholders
The Hands-on Leader
This should be an obvious one. With Github, Quora, Slack and many other communities to help data scientists with independent learning, data scientists have great respect and appreciation for managers with solid technical background.
Being able to roll up your sleeves to develop algorithm from scratch, or challenging your data scientists when their chosen machine learning techniques don’t work in production environment because the underlying assumptions don’t hold, hands-on leaders are highly respected within strong, technical teams.
The Thinker – Focus on Good Problem Solving Approach
Sound, problem solving approach to break down the problem, view it from another angle and achieve breakthroughs.
Many a times, data scientists need to connect the dots which are seemingly unrelated and they need certain inputs. Sometimes they need granular data which are not available, sometimes they need certain talent which is unavailable. Once I met a hiring manager who wanted to hire a senior data scientist with experience in computational modelling and human cognitive behaviour. Back then, it was almost impossible to hire someone with this combination of skill set.
A good leader with a sound, problem solving approach is able to break down the problem, view it from another angle and achieve breakthroughs. Data scientists tell me that such managers are ‘really smart’, and there is so much to learn from them.
Building a Legacy
Another observation is that successful data science leaders have a loyal team. Data scientists follow this leader wherever he/she goes, and this is a strong leverage for Chief Data Scientists who are building teams from scratch. You know you have access to good talent anytime.
A handful of successful Chief Data Scientists have the tendency to ‘put themselves out of a job’. They groom their data scientists to a point where the team functions independently with minimal guidance. These managers, they build an institution, they build a legacy that continues even when he leaves the organisation.
Visionary managers tend to take succession planning into their own hands. They groom future leaders, impart skills and most importantly, impart knowledge and wisdom. I have placed mid-level data scientists into organisations and witnessed them blossom into outstanding data science heads not just within the firm, but also across the industry. They tell me “I need to thank my manager for giving me those opportunities”.
Candidates often ask “what does it take for one to take on a Head of Data Science role?”
Well, the answer would be highly dependent on the mandate of this role but I’d say – definitely a good head above the shoulders and a big, strong heart. A big heart to impart that knowledge you have painstakingly acquired over the years, and yet have to give away so easily. A strong heart because implementing large scale analytics solutions can sometimes be a marathon with the end point nowhere in sight.
Do you have what it takes to be an effective data science leader?
Su Yi has 6 years of recruitment experience across the data science, technology and financial services sectors. She has worked with MNCs, start-ups and statutory boards across the financial services, technology, healthcare, retail & consumer, media and public sectors. Her achievements include making successful placements from mid to C-level hires across Asia Pacific, including Singapore, China, Malaysia and South Korea.
