In our experience there is no doubt that Data Scientists are one of the most in-demand professionals of the 21st century when it comes to hiring. We all remember the Harvard Business Review in 2012 naming “data scientist” as the sexiest job of the 21st century, but since then the hype has grown exponentially. With this hype comes an array of issues for businesses looking to hire them.
Simply by posting a job advertisement online and receiving hundreds of applications, one would be forgiven for thinking it’s easy to hire a Data Scientist. Here stems the first problem; everyone wants to be a Data Scientist. Ok, “everyone” is an exaggeration, but there are certainly a lot of people from a mix of backgrounds who hear of six figure salaries and weird and wonderful artificial intelligence or deep learning projects in data science and want to make the transition. We think the terminology of “fake Data Scientists” is a step too far - there are very transferable skill sets between professions, but the definition of a Data Scientist is certainly becoming more indistinct. New job titles are emerging for very clearly defined roles such as Machine Learning Engineer, Computer Vision Specialist, NLP Scientist. Companies need to be able to read beyond a job title and get a good understanding of a person’s CV and come to the realisation of problem number two, there is still a huge shortage of talent.
The hype over the last few years has led to a lot of people making the transition from academia to data science. “Data Analysts” are going back to University to study bridging degrees (UCL’s Computational Statistics and Machine Learning Masters programme must be one of the most oversubscribed in the UK), the emergence of dozens of “data science boot camps”, and a huge uptake of online courses such as Coursera. However, this increase in supply nowhere near meets the huge increase in demand. Data Science is not restricted to Silicon Valley-esque tech businesses; this year alone Xcede have recruited Data Scientists for law firms, investment banks, sports teams, charities, and even pet-care businesses… the list of industries and their need for Data Scientists is everlasting. This Harvard Business Review feature reported that "companies in the top third of their industry in the use of data-driven decision making were, on average, 5% more productive and 6% more profitable than their competitors". An attractive part of the role of a Data Scientist is that it is possible to very clearly quantify the added value to a business: a contributing factor to the huge increase in demand.
Basic economics dictates that high demand and low supply leads to an increase in price – problem number three. Data Science salaries are rather mixed and inconclusive at the moment. Six figure salaries for little commercial experience are just as common as £50,000 “Head of Data Science” roles. A lot of this stems from problem number one; defining the role of the Data Scientist. Genuine “unicorn” Data Scientists are likely to have a host of offers when changing jobs which can lead to a bidding war between businesses. A maturing freelance/contract market is also adding a spanner to the works with lucrative day rates blowing salaries out of the water.
This increase in demand leads to another problem: problem number four – how to best attract and interview Data Scientists. Xcede are one of many digital agencies that recruit for the Data market in the UK. Data Scientists get dozens of LinkedIn messages each week, making it harder and harder for companies to stand out and attract the best talent. LinkedIn messaging is not the only way to attract Data Scientists! However, once you get people through the door, you then have to try and work out how to interview them. It is a fine art producing an interview process that allows you to: test someone’s skill set but still be able to sell the role, opportunity and company; have a fast-paced process so that you don’t miss the best talent who naturally get appointed first. While ensuring it is not too quick that it is a rushed hire and have a process that includes some test of technical ability, but not one that requires such a commitment from a candidate that it can put people off, e.g. take-home tasks that take a whole weekend to complete.
It is fascinating where the Data Science market is heading, with the job title itself potentially dissolving into many other more clearly defined job titles. This may help balance fluctuating salaries and allow companies to easier attract and interview people, whilst making sure both companies and candidates have a much clearer representation of what the role will involve and what the candidate’s skill set is.
Some advice to consider:
1. When reading applications, read beyond the job titles.
2. Make sure you carefully benchmark your salaries, and don’t enter unnecessary bidding wars – the right people will want to work to you for the right reasons, not just based on salary.
3. Understand that interviews are a two-way process – make sure you tell people why it’s great to work for you!
4. Put together an interview process that doesn’t require excessive commitment from candidates and one that doesn’t drag on too long.