Second only to GDPR, recruitment may be the biggest challenge facing data and analytics functions. This is partly a function of high levels of employment across the UK and partly a reflection of historically low levels of graduates from science, technology, engineering and mathematics courses.
It is also a result of the novelty of data and analytics as a career path. If you started your degree in 2010 and then went on to take a PhD, demand for your skills set as a data scientist, for example, really only took off two or three years back. But take off it certainly has with demand outpacing supply - according to jobs board Indeed, there were 2.4 open positions for data scientists for every one candidate.
Over-demand and under-supply of employees always leads to wage inflation, adding another pressure on the ability of growing data and analytics practices to staff up. So is the answer to be more diverse in hiring profiles, adopting new recruitment techniques to broaden the candidate pool, or to press universities to accelerate their throughput of graduates?
To find out, DataIQ Summit this year featured a panel, chaired by Toni Sekinah, research analyst and features editor, which asked three leaders in the industry about their own strategies and experiences. All three agreed on one thing - that the best hires are not just individuals with a check list of technical skills.
“Young people with curiosity will have a talent for data.”
“We need people to be curious about data, to understand it, to be technically capable and to come up with strategies to drive change,“ said Judith kleine Holthaus, senior analytics and insight manager at Whitbread. She believes that the supply wil become more abundant as a result of the digitisation of all aspects of our lives: “Young people with curiosity will have a talent for data.”
To spot this disposition, Whitbread uses problem-solving tests during its recruitment process to understand how a candidate thinks about data and how it can be applied to business. But it has also started to offer internships for people who already have a grounding in the basics.
Santander UK has taken this a step further. “We have an apprenticeship scheme which is in its second year. It gives us a pipeline of very young people who are learning skills and about the bank at the same time. That gives them the ability to explain what data means in the organisation and the technical skills that are useful. We are teaching them how we want them to be and allowing them to thrive,” explained chief data officer David Hayes.
The organisation faces a specific recruitment issue by having its data function based in Milton Keynes, which is close enough to London to be a commuter town but where the salaries on offer don’t match those available in the capital. “We are able to recruit financial services, risk and IT engineers, but not new data skills sets. To recruit data science, we are starting to look in London and we need to partner with universities, but there aren’t any in Milton Keynes,” said Hayes.
Partnering with tertiary education establishments is becoming more common among data-driven organisations who want to identify potential data scientists early in their PhDs or Masters. “We have partnerships with six universities and we are trying to get involved with their curricula through one of our data scientists who is an external professor,” said Wade Munsie, director of advanced analytics and BI development at Royal Mail.
His function is in the process of doubling in size from 30 practitioners, so finding the right candidates is a very pressing issue. Munsie is tackling this in quarterly bursts and has been looking to use alternative methods from conventional recruitment agencies which he finds often do not have the depth of understanding which they claim.
“Data engineers now are where data scientists were two years ago.”
One solution he used was to run a hackathon which was promoted through LinkedIn. Out of 70,000 views, 60 applications were submitted, leading to 30 interviews and 16 candidates who spent the day working on a specially-created data set and problem. Munsie was able to recruit three junior data scientists right out of this pool.
His biggest recruitment now is to find more practical, technical skills. “Data engineers now are where data scientists were two years ago,” he said. Hayes expressed similar difficulties: “It is hard to get in good data management people, especially those who can deliver something of value to the organisation and make it work. Those are rare skills which obviously make recruitment difficult. We don’t need a lot of them, just more than we have.”
All three acknowledged the impact which wage inflation has been having, most obviously in the salary expectations of data scientists. “If you overpay people, unless they recognise that in the value they are creating, it will unravel. The world of data continues to grow which will hit recruitment because more companies are competing. Data is moving up the hype curve - I hope it stays at the top for some time - but we won’t justify that hype without showing the added value for the organisation,” said Hayes.
“Everybody needs to avoid over-paying and out-bidding each other.”
Diversity has been one of the key approaches by data and analytics recruiters for a number of years, looking outside of the STEM pool for individuals who have the right mindset and can be trained in specific technical skills. According to kleine Holthaus, “we have a passion for diversity - we don’t just want people with a university and quantitative background, we need different viewpoints. We need people to bring to life STEM and problem-solving in the curriculum. I would love to see more of that.”
Like any growth sector, recruiters face the dual problems of limited supply and over-claiming by candidates keen to get in on the action. The three brands represented at DataIQ Summit are conscious of both problems, but also see industry-level solutions as being important.
As Munsie pointed out, some of the fixes could come from a degree of self-restraint and mutual self-interest. He said: “We all contribute to the supply problems because some companies are paying over the odds. We don’t pay six figures for a mid-level data scientist. Everybody needs to avoid over-paying and out-bidding each other.”