Companies that embrace diversity and especially women, embedding their data science team will outperform competitors, says research
According to research by The Alan Turing Institute, women make up just 20 per cent of AI and data professionals, despite accounting for half the UK population.
The reasons for this are many and varied but it’s a trait that is hugely damaging to the sector and its potential to add value to businesses and society more generally.
‘Technology leaders report that diverse teams offer better innovation, collaboration and business engagement’
“All of our research tells us that diverse teams simply provide better results,” says Theadora Norris, data and business intelligence senior consultant at recruiter Harvey Nash. “Technology leaders report that diverse teams offer better innovation, collaboration and business engagement, among a host of other metrics.
“Vital decisions are made on the basis of data every day, whether that’s in corporate businesses, healthcare or government, so it’s really important that the teams that are collecting, managing and exploiting this data are representative of the populations they serve so there aren’t blind spots in the work that they do.”
Having a gender-diverse team means a greater understanding of potential problems that need to be solved will be brought to any discussion, says David Field, head of talent at Speechmatics. “Speech recognition datasets, for example, are historically made up of middle-aged white men without speech impediments or whose accent is the native one for the language being spoken. A more gender-diverse team is likely to flag this issue before the tech is available for the wider public to use, who would surely take issue with a product or service that doesn’t work properly for over half of the British population.”
Why having women on your data science team is important
Virgile Raingeard is co-founder of compensation benchmark platform Figures. He gives the example of a high-profile case that demonstrates the importance of having a diverse team. “A standout example of AI bias negatively impacting women was in 2018 was when Amazon’s hiring programme ruled that male candidates were preferable, and put them forward more regularly for positions,” he says. “This was because the training dataset itself was skewed towards this conclusion by containing a greater number of CVs from males. A gender balance would have made it more likely such skewed data was spotted and rectified.”
There are other, more general reasons why this matters.
Karin Sasaki, senior consultant in data science at Ekimetrics – who works in a team that is split 60:40 in favour of women – points out that people with different backgrounds, experience and expertise will bring different perspectives. “I believe this variation enriches a workplace and helps teams perform better, especially on a technical level,” she says. “Additionally, being part of a diverse workplace naturally makes everyone more tolerant of people’s idiosyncrasies.”
Women in data science
Saurabh Upadhyay, chief people officer at data science firm Tredence, believes a gender-diverse team can help create a more positive work environment. “It can also help to attract and retain more talented employees,” he says. “Finally, a more diverse team can help an organisation become more innovative and competitive, while a lack of diversity can lead to a lack of understanding of the needs of female customers and employees.”
There are steps firms can take to improve the diversity make-up of their data science teams.
Mary Shea, global innovation evangelist at software firm Outreach, says companies need to improve their hiring practices. “The language used in job descriptions can have a big impact on who actually decides to apply,” she says.
“Using more inclusive language encourages candidates from all types of backgrounds to explore the job opportunity. Interview panels also need to have diverse representation. If you want to hire more women, then your interview loops need to include more women. Most employees don’t want to be the lone representative for a specific group at their company.”
Talent pipeline
Fei Sha leads the data science and machine learning department, which is 80 per cent female, at career development platform Degreed. She believes the issue goes deeper.
“We cannot simply rely on hiring managers and recruiters to source talent because it’s too late for that,” she says. “Instead, you need to look further up your talent pipeline to the generations of girls who are currently in school and college or university. Developing their interest in STEM fields now will pay off for hiring managers who will be looking for data scientists in the next decade.”
There are signs this is starting to improve, believes Norris: “Industry research suggests that starting to learn code at age 10 can help encourage more girls down the career path, and from there we need even more role-modelling, apprenticeships and commitment from businesses to train women and girls into this rewarding career.”
Female role models
Having relatable role models in senior positions is also important. “Businesses can attract more women into the sector by ensuring that senior leadership team members are not just men,” says Field. “That way, women can see that they’re being represented at all levels. Making sure women have mentors and advocates within the business is another way in which businesses can attract women to join the field of data science.”
It’s also important that organisations tackle any issues that may prevent people from staying in the sector. “Companies need to ensure they have created a culture of inclusion and belonging for women such as flexible working policies, equal pay for equal work, frequent anonymous employee wellbeing surveys, and visibility and support on career progression,” says Raingeard.
Figures’ own research suggests that tech companies that are open with their staff about how much each person earns have fewer – and in some cases non-existent – gender pay discrepancies, he adds.
Battle for the best talent
Not only does having a more diverse pool of data scientists make sense for the reasons outlined above, the bottom line is that with such talent increasingly in demand, organisations can ill-afford to ignore women.
“The battle to acquire the best talent has already begun,” points out Shea. “As companies of all shapes and sizes prioritise profitable growth versus growth at all costs this will only intensify. Companies that wholeheartedly embrace diversity and embed it into all of their processes will outperform their competitors.”
More on diversity & inclusion
Why diversity matters when recruiting cybersecurity staff – Putting diversity at the heart of your cybersecurity team helps you spot issues and problems that might not have occurred to you
Using artificial intelligence to promote diversity & inclusion – Artificial intelligence can help remove unconscious bias when recruiting to fill tech positions. But can it be a double-edged sword for tech leaders when it comes to promoting diversity & inclusion?
See also: How to embark on a data science career – the key factors to consider
