As we look back on 2019, we’re taking a look at the trends and making predictions for what companies who plan to hire data scientists in 2020 will be facing. (7 minute read)
It’s Official – Uncle Sam is Hiring Data Scientists in 2020
“Hiring Data Scientists Isn’t an Exact Science” was the recent headline in an article that highlighted the fact that in mid-2019, the US government’s Office of Personnel Management (OPM) finally recognized the job title “Data Scientist” as an official role in government agencies. (If you are interested, here’s a link to how it officially describes the role.)
In doing so, the OPM gave agencies around the nation the green light and budget to create job descriptions and go out and hire data scientists “officially”.
As we all know the government is usually the last to jump on board any cutting-edge trend.
So, what does this anecdotal story signal if you are planning to hire data scientists in 2020?
If the ginormous US government is starting to hire data scientists in 2020 that’s HUGE. It means that the role of a data scientist is in great demand. Which brings us to our first trend for 2020…
Companies Plan to Hire More Data Scientists in 2020
In 2019, some speculated that the data science job market would be flooded with so many undergraduates and bootcamp-certified or Udemy-trained data scientists that the job market would soften. Well, those guys and gals were wrong.
According to LinkedIn’s recently published 2019 Emerging Jobs Report
“Data science is booming and starting to replace legacy roles.“
Companies are hiring data scientists for new roles and new jobs. There’s a long list of evidence to support this.
Starting with the aforementioned Feds, who also recently claimed they need more employees for cybersecurity, data science, AI, robotics automation and machine learning. Agencies such as the Homeland Security and the Office of Management and Budget have identified critical vacancies in these fields to be filled in 2020.
Other companies are planning to hire more data scientists too.
One sure signal of this is that analytics recruiters such as BurtchWorks and Harnham are expanding their analytics recruitment teams.
In fact BurtchWorks found that 82% of clients they surveyed in 2019 were planning to hire data scientists in the first half of 2019 and 70% were actively hiring data scientists to grow their teams.
One LinkedIn report cited that data science job openings increased 60% in 2019. The Bureau of Labor Statistics also predicted that between 2018 and 2028 the job will grow by 16%-30%.
Perhaps the real acid test of a robust job market is campus recruiting. Higher education institutions who are churning out data scientists are finding that those graduates are getting jobs easily.
Basically, any study you look at will show that data science job numbers grew in 2019 and show no signs of slowing down.
One reason companies plan to hire more data scientists is because there’s a lot of room for growth in data science and analytics across a wide variety of industries. As you can see from the image below (courtesy of BurtchWorks), data science has only just begun to expand as a discipline into industries other than technology (which also has room for growth). Notice the US Government is last!
To sum up, data scientist hiring won’t slow down in 2020.
Data Science Skills Will Still Be in Shortage
Some optimistic research firms like Gartner predicted several years ago that by 2020 whole organizations would be data savvy and automated enough that the data science skill shortage would be alleviated.
Alas, that’s just not the case yet. For starters, higher education has not caught up to demand. There is still an apparent shortage on data literacy that we’ve yet to fix. A Gallup study predicted that by 2021, 69% of employers expected that candidates with data science skills would be given preference for jobs in their organizations. Yet only 23% of college and university leaders say their graduates will have those skills by 2021. That’s a pretty big gap!
There are plenty of other indicators that data skills are still in relatively low supply. Coursera conducted its first ever Global Skills Index survey in 2019. It determined that the US is lacking in data science skills, ranking at number 16 globally. Yikes.
BurtchWorks seconds this notion, they’ve observed a clear trend of companies importing data scientists from China and India to fill skills and experience gaps.
KPMG’s 2019 CIO Survey asked the question “Which functions do you feel suffer from a skills shortage?” The responses are depicted below. Top of the list? You guessed it!
Global companies take note: Hiring a data scientist outside of the US will be no easier either.
Outside of the US, the talent gap is clear. Europe needs 346,000 more data scientists by 2020. Demand for data skilled workers in the UK has more than tripled over the last five years to 231%. According to O’Reilly Media almost half of all European companies are struggling to fill their data scientist positions.
The reasons for this prolonged skills shortage are many, varied and at times unclear. Until they are identified and addressed, its unlikely much will change in the US market or elsewhere for data science skills.
Hiring (and keeping) Data Scientists Will Still Be a Pricey Endeavor (for now)
In 2015, according to BurningGlass the average salary for a data science and analytics professionals was around $93,000. While many entry level data science roles may still start pay towards that end of the spectrum or less, the average salary has increased steadily since then.
According to the Bureau of Labor Statistics the average salary of a data scientist is now around $118,000. This statistic is supported by many other sources as well, such as Harnham and Randstadt’s 2019 Salary Surveys, which found average starting salaries for data scientists to be around $105,000 – $121,000. Specialists such as NLP or AI engineers earn even higher averages salaries.
To add insult to injury, according to Harnham two thirds of data scientists receive on average about a 20% bonus on top of their salary. Whoa. That’s expensive.
Further exacerbating the data science salary issue for employers is the high level of turnover and the willingness of many data scientist to jump ship at any moment after negotiating a higher salary somewhere else. According to Harnham, a discouraging 74% of data scientists would leave their role if the right opportunity came up.
This is reflected in the fact that data scientists on average in 2019 stayed in a role for just 13/4 years. The average increase in salary sought by data scientists looking to change jobs was 25%, while the average increase they realized was 14%.
The good news is that we’ve learned through our QuantCrunch Report research and university data challenges that many universities are starting to churn out talented undergraduates with freshly minted data science degrees. Many of these students will have gained significant work experience while in school by working on real life data projects in class, through internships, data science club projects and data competitions.
As their numbers increase and companies begin to realize that these graduates can bear some of the load of more expensive data scientists, we may see some salary pressure alleviated. This would be good news for the many organizations and data-driven startups who have fewer resources, or who are low on the data maturity curve and who need to start building data science teams and capabilities.
AI, Cloud, IoT, and Machine Learning Will Drive the Need to Hire Data Scientists
Looking at US-specific data presented in the table below from the World Economic Forum’s Future of Jobs Report in 2018, 89% of US-based companies are planning to adopt user and entity big data analytics by 2022. More than 70% want to integrate the internet of things and take advantage of machine learning and cloud computing.
|Planned technology adoption by US companies within the next 4 years
||Share of companies surveyed
|User and entity big data analytics
|Internet of things
|App- and web-enabled markets
|Augmented and virtual reality
|Distributed ledger (blockchain)
Other indicators of explosive growth in these areas abounds. Jeff Dean, Senior VP at Google AI, recently Tweeted the graph to the right showing that approximately 100 Machine Learning research papers were being published daily at the end of 2018.
And the National Venture Capital Association says that 965 AI-related companies in the U.S. have raised $13.5 billion in venture capital through the first 9 months of 2019.
With big plans for investing in all sort of new technology and the rising interest in machine learning, it’s no wonder that according to LinkedIn hiring of AI and Machine Learning engineers has grown 74% over the last 4 years.
As continued rapid advancements in AI and machine learning enable data scientists to do even more with their craft there will be a steady need for people to develop, direct and monitor these highly advanced programs.
It’s worth pausing here to introduce the next four trends with a summary:
In 2020 the nature of the data science workforce and skill make-up will need to steadily adapt to meet the talent shortage.
In 2020 Companies Will Hire Data Scientists from Within by Reskilling and Upskilling
So back to the Feds. They may be slow to acknowledge that they need to hire data scientists and the last industry to recognize the job role, but after doing so, right away the government identified that its agencies will need to reskill employees and hire from within to fill the critical roles previously mentioned. They’re starting, very appropriately, with Census Bureau employees.
One driver of the rise in reskilling, besides the data science skill shortage, is the growing recognition of the need for domain expertise for data science strategies to be successful. A recent article featuring data-driven startup Zonehaven mentioned that the company would soon be looking for data experts who have experience in public safety described as,
“someone who has a computer science degree but also worked as an EMT in college”.
That mix of skills and knowledge is no doubt hard to find. For data science and especially artificial intelligence applications to be useful in the real world, solutions developers need real domain expertise and insight. Since data science is a fairly new profession, and machine learning even more so, it’s still hard to find people who have both experience in data science and domain expertise.
McKinsey has thus concluded recently that companies need to reskill and upskill their existing employees to both meet the skill shortage and educate business professionals who have domain expertise by establishing internal educational programs aimed at reskilling. As they put it, s
“As organizations rebuild their foundations to compete in the era of data and advanced analytics, in-house capability building programs offer the best way to train workers up to the task”.
In order to exploit the potential of both data science and AI, staff and end users will need to be able to think and work in terms of data and analytics, and learn how to embrace data exploration. That requires education.
Hiring data science talent externally has been the primary way companies have addressed pressing human resource needs in the field. This tactic is unsustainable as the current skill shortage, high salaries and high turnover demonstrates. To meet the rapidly growing need for a wide and deep variety of data and analytics skills and to create a true data driven organization at scale, companies will need to train and hire existing employees at all levels to work in the realm of data.
Executives and Managers Will Go to Data Science School
In terms of upskilling, business leaders should be targeted in 2020. The data talent problem does not only lie with practicing data scientists. It also extends to existing job roles from the C-suite to the business frontlines — all of which are increasingly enabled by analytics and data science.
Why upskill executives? Companies hire data scientists because they want to develop data-driven strategies to remain competitive. They say they want these data scientists to use terabytes of data produced daily to generate new strategic insights and revenue streams and to solve business problems.
Yet after years of hiring teams of data scientists to do this, most company’s fail in their data science efforts (85% according to Gartner).
The culprit? We’re looking at you C-suite. In Harnham’s 2019 salary survey, “poor management” was cited as the number 2 reason that data scientists leave companies. It could be argued that the number 1 reason for turnover – lack of career progression – could also be attributed to the feeling that data scientists do not have opportunities to be taken seriously in many companies.
Numerous studies and anecdotal musings from disgruntled data scientists on Towards Data Science indicate that data science projects fail because they do not get the support and/or understanding of their value from business leaders.
Why is this? Business leaders and other consumers of data science projects still lack understanding of the value and meaning of data science insights. Part of this issue is a lack of so-called “data translators”. Additionally becoming “data-driven” requires leadership insistence that business managers stop relying on gut and more on data.
The solution to these issues going forward is to educate leaders on data science concepts, choose leaders to act at data translators within the business units, and then, once everybody’s on the same page make it clear to all stakeholders that decisions will be data-driven.
And if you need some convincing to do this, take a look at the graph to the right which depicts the difference in analytics education at high performing companies per McKinsey’s analysis. The gap at the executive and manager level is HUGE.
Plan to Hire a Diversity of Data Scientists for Your Teams
It’s no secret that there is societal pressure to increase the number of women and minorities in the field of data science and tech in general. There’s much talk and research indicating a lack of diversity in data science including a recent diversity report by Harnham which found that in the US, UK and Europe gender pay gaps and seniority gaps still persist in the field of data science. The report cites that,
“the fact that Data teams are more cognizant than ever about the need to make significant changes, only 23% of roles are currently filled by female professionals, largely a result of low representation in Data Engineering and Data Science.”
Both Harnham and Burtchworks 2019 data science salary studies show that while women have made strides at entry level data science jobs, approaching between 25%-30% of the data science profession at the junior level, their numbers fall dramatically with a rise in seniority to about half that of men at higher levels of management.
The numbers are still pretty bleak for ethnic minorities in data science as well. They only represent about 12%-15% of the data scientist population.
With McKinsey and Co. and Dow Jones VentureSource churning out studies that show that companies and tech startups with diverse teams perform 15%-35% better than those without diversity you can expect more companies to start re-thinking their diversity and inclusion policies and hiring diversity managers to address this talent issue in data science.
Hiring Data Scientists out of Bachelor Degree Programs Will Become the Norm
Our conversations with numerous universities, as well as the profiles of students participating in our data challenges indicate that in addition to developing new data science degree programs, universities are putting a real focus on incorporating data and analytics skills into the wider curriculum.
As professors of data science recognize that hands-on learning with messy data is what undergrad students need and as they ramp up the real world skills like Python and SQL taught in these newer degree programs, companies are realizing that they can test out undergrads with their own in-class data projects and internships. They can then look to hire these entry level graduates and develop them internally into home grown data scientist.
We’ve certainly had clients tell us that they are moving away from hiring experienced data scientists who are hard to retain and hiring less experienced graduates that can be trained in house and hopefully better retained as a result.
Finally, some universities are seeing a movement of enrollments away from the humanities and traditional quantitative majors such as actuarial science and into data science and data analytics degrees. More students in non-quantitative majors are enrolling in electives that are data science focused. This trend will no doubt mean that companies will see undergraduates as we way to hire data literate employees and develop data citizens.
What to Do in 2020 When it Comes to Hiring a Data Scientist?
Despite speculation in the last year that the job market for data scientists will soften or that the shortage of skills will be alleviated through undergraduate programs, bootcamps and reskilling, in 2020 it still will be challenging to hire and retain a data scientist.
The data science and analytics job market will remain candidate-led in 2020. Competition for talent will be stiff, and perhaps rapidly growing as new AI-driven companies and corporate AI centers of excellence come onto the scene.
It’s no longer enough to just hire a data scientist or a few of them to come in and make sense of big data and AI.
Organizations will have to take steps to counteract the ongoing challenging market forces through a variety of measures. These include getting serious about creating a data-driven culture at the top levels, educating and upskilling business leaders and employees to become data literate and stewards of data, and creating career paths in data science departments that promote diversity and inclusion as well as challenge and opportunity to grow and develop.
These tactics will be especially important for companies and startups who cannot necessarily afford to pay the highest salaries to attract external talent.
One thing’s for sure, whether you are a company or a candidate, 2020 is sure to be an exciting time of change, growth and progress in data science!