Back To Top
DataEthics4All-AI-DIET-Insights-Hero

Guide: Debunking the Diversity & Inclusion Pipeline Problem in Tech

DataEthics4All-Circular-Logo

Guide: Debunking the Diversity & Inclusion Pipeline Problem in Tech

RESEARCHERS
Quentin J Louis, Joshua Walker, Katherine Lai, Victoria Matthews, Kitty Atwood

PRINCIPAL INVESTIGATOR
Susanna Raj

DESIGNER
Ayomikun Bolaji

Contents.

Executive Summary.

INSIGHTS.

Many company pledges have failed to redress this problem

This is not because of the ‘pipeline problem’

Instead, companies choose to hire according to decontextualized metrics that reinforce accumulated privileges. This ‘decontextualised meritocracy’ reinforces the status quo and leads to little hiring from non-elite institutions.

The technology sector remains lacking in diversity.

This is despite powerful incentives to become more diverse and inclusive.

Diversity alone is not enough; inclusion is needed

To boost both, the commitment of leadership to a clear strategy is vital

This strategy must include using survey data to assess how well managers are fostering inclusion; this data can be used to identify working strategies.

Human diversity abstract concept vector illustration set. Nationality, race and ethnicity, country of birth, passport, social difference, human rights, skin color, genetic code abstract metaphor.
DataEthics4All Membership Diversity check points

Diversify searches for talent beyond traditional, elite institutions.

DataEthics4All Membership Diversity check points

Focus on retaining talent.

TOP 2 PROPOSALS.

Introduction.

In this report, we shall establish the state of diversity and inclusion in tech, then seek to explain why it is stagnant.

We will show that companies are incentivised to foster diversity and inclusion, and that they are able to do so. However, a preoccupation with established pathways to the tech sector have led to ‘mirrortocracies’ which harm companies’ business prospects. To counter this, we recommend that companies broaden their horizons on recruitment to tap into a wider range of diverse talent.

On inclusion, as a first step company leadership should draft a clear and enforceable strategy and commit to carrying it out. The specifics of such a strategy are best left to managers working on the ground level, but through surveys of employees, the strategies of these managers can be assessed and replicated or replaced as appropriate.

THE ROLE OF THE COMPANY IS THEREFORE TWOFOLD:

To force managers to work towards inclusion.

To establish through data the best ways of doing so.

DIVERSITY REFERS TO VARIETY IN THE MAKEUP OF GROUP, ON:

Gender and sexual orientation

Age

Social Economic Background

Ethnicity

Disability

Diversity and Inclusion are not the same thing.
Inclusion is the extent to which the people in this group feel part of the group – more difficult to measure, but vital.

Diversity is important in tech (and elsewhere) because it brings different perspectives and backgrounds to problems, increasing the chances that a solution can be found.

Inclusion is vital as only when people feel included will they offer their unique perspectives without fear of ridicule or exclusion.

This, as we will see, is reflected in the data, where more inclusive companies benefit commercially.

Background.

Diversity and inclusion in the technology sector have been historically lacking, and improvement has been slow or non-existent. Women in particular are underrepresented, having made up around 17% of the tech and IT workforce for the past decade and only 20% of executive boards in the UK and US in 2019.
(Understanding the UK AI labour market: 2020).

On ethnic diversity, minorities are better represented –  in the UK, ethnic minorities make up 16% of the tech workforce, compared to 13% of the national population.
(The voices of our industry BIMA Tech Inclusion & Diversity Report 2019). 

However, this masks two key points:

First, that Asians tend to be over-represented (to the point where Asian men are not considered ‘diverse’ when recruiting for tech; Understanding the UK AI labour market: 2020). whilst other minorities tend to be under-represented.

Second that diversity in the workforce does not automatically mean equality or inclusion. This can be seen in the lack of diversity in leading positions in tech companies – for example, in the USA less than 10% of tech entrepreneurs were women between 1990-2016 and as can be seen below, ethnic minorities are just as underrepresented. (Diversity in Innovation, Cambridge 2017).

US Population (2019)

[visualizer id="13697" lazy="no" class=""]

% of Entrepreneurs by Race (2015)

[visualizer id="13694" lazy="no" class=""]

Moreover, women and minorities are more likely to report discrimination35% of women in tech said their career progression had been discriminated against due to their gender, compared to 10% of men. 40% of black respondents and 31% of Asians said that ethnicity was a factor, compared to 9% of white respondents. 34% of 55-64 year olds said they suffered discrimination due to their age, significantly higher than any other age bracket. (The voices of our industry BIMA Tech Inclusion & Diversity Report 2019). 

Discrimination and unfairness leads to an exodus of affected workers; unfairness was cited as the most common reason (37%) for voluntary exit from tech jobs. Women and minorities were significantly more likely than men and white workers to experience unfairness and leave because of it. 

61

% SAW THEIR COMPANY’S INCLUSION AS NEGATIVE

Ethnicity and gender are not the only areas of concern.

Age has already been mentioned but sexual orientation (LGBT employees are most likely to be bullied – 20%; Scott, 4) and geographical location (15 cities containing 31.3% of the US population receive 94% of venture capital investment) also matter. (‘Tech Clusters’ Journal of Economic Perspectives 34.3).

Finally, it is clear that diversity does not necessarily create inclusion – in an analysis of tech employees’ job reviews, 52% expressed positive sentiment about their company’s diversity compared to 31% who assessed it as negative. However, only 29% saw their company’s inclusion as positive, 61% as negative. (Dahbi 34).

What this outline has demonstrated is that the technology sector has serious problems with diversity, discrimination and inclusion of many different groups. What follows will seek to explain why this is and what companies can do to reverse this trend.

Incentives.

PWC-logo-DataEthics4All-A-Short-Guide-to-Diversity-and-Inclusion-in-Tech-
Kapor-Center-Logo-DataEthics4All A Short Guide to Diversity and Inclusion in Tech

Tech companies have many incentives to encourage diversity and inclusion. The most important according to companies themselves is to attract and retain talent, which 47% of tech companies listed as the primary objective of diversity and inclusion initiatives (see below) (PWC: Diversity & Inclusion Benchmarking Survey).

As described above, turnover is driven in no small part by discrimination and unfairness. Conservative estimates suggest that the tech industry loses $16bn/year on re-hiring due to employees quitting because of toxic work cultures.

57% of leavers said they would have stayed if their employer had taken steps to make the work environment fairer and more inclusive. In addition, companies which had diversity and inclusion initiatives in place were significantly less likely to lose workers due to feelings of unfairness or mistreatment. (Kapor Center: Tech Leavers Study, 2017).

This shows that companies have incentives to pursue inclusive policies to retain  workers.

In addition, substantial evidence suggests that diversity and inclusion are strongly correlated with business success. Companies which ranked in the top 25% on gender diversity on executive boards were 25% more likely to be more profitable than those ranked in the bottom 25%. On ethnic diversity, the difference was 36%.

Moreover, this gap is widening over time (the gender diversity profitability gap was 15% in 2014), which suggests that less diverse companies are falling further behind. (Diversity wins: How inclusion matters, 2020).

Tech companies are particularly incentivised by ‘neurodiversity’ – mental conditions such as autism – as they value employees who think in a non-linear, pattern-forming, detail-focused or big-picture way, all of which have been associated with autism. However, 39% of neurodivergent people hide their condition, in part due to fear of negative reactions that come from exclusive work environments. (The voices of our industry: BIMA Tech Inclusion & Diversity Report 2019). 

[visualizer id="13971" lazy="no" class=""]

Diversity has been shown to influence employee engagement: a 2005 Gallup study found that amongst the top quartile of employers who thought their company had a strong diversity focus, only 1% were actively disengaged. Amongst the bottom quartile, 38% were.
(‘From Diversity to Inclusion: An Inclusion Equation’, in Diversity at Work: The Practice of Inclusion).

This shows that companies have incentives to pursue diversity and inclusion – to attract and retain talent and improve business performance.

The question remains: 

why do many not act upon these incentives, or if they do,

WHY DO THEY FAIL?

Company Pledges.

Although there are clear incentives for companies to diversify their workplaces, there appears to be a general reluctance to do so. Major tech companies began to publish annual diversity reports in 2014, but few have made any substantial, let alone meaningful, ground in their stated goal of hiring more ethnic minorities.

Many companies (including Apple, Facebook, Microsoft and Twitter) have seen low single-digit increases in their percentage of Black employees since 2014. So, despite acknowledging the gap in their workforces with a public pledge to increase diversity, this has not come to fruition.

In response to the Black Lives Matter protests and the global racial justice movement in 2020, tech companies have been forced to become increasingly aware of the role they play and their impact on wider society.

BLM-Symbol-DataEthics4All-A-Short-Fuide-to-Diversity-and-Inclusion-in-Tech

Out of the 240 tech companies analysed in the Blendoor State of DEI in Tech 2021 report, 159 companies made over 535 DEI pledges in 2020 with a total cash value of $4.56bn. However, there has been little evidence of tangible progress overall. Less than half of the tech companies analysed in the report made a statement or pledge concerning Black Lives Matter, but only 70% of those companies acted to publicly disclose the percentage of Black employees in their workforce.

Financial pledges and monetary displays of support are not the same as pledges to hire more diversely and to retain those employees. It should be taken into consideration that financial pledges are possibly a ‘quick fix’, especially for cash-plenty companies, rather than challenging the very internal issues and biases within their companies.

Debunking the “Pipeline Problem“.

In 2016, Facebook blamed their failure to hire more people of colour on the public education system. This caused a stir and ignited the discussion of the “pipeline problem”.

This suggests that because people have not had the opportunities to access certain areas (such as studying computer science in public high schools), they will inevitably ‘lose out’ to those that do. Thus, the solution lies outside of the company – it is out of the company’s hands.

But the lack of diversity in a company’s workforce is not a “pipeline problem”. This is a myth perpetuated by companies so they do not have to look at the biases in their hiring processes.

It is untrue that there is a limited pool of talent in underrepresented groups to recruit from. In 2014, top universities graduated Black and Latinx computer science and engineering students at twice the rate leading tech companies were hiring them.

A gulf between Silicon Valley and top graduates at U.S colleges and universities emerges.

Employees of top 75 Silicon Valley tech firms (gender, %)

[visualizer id="13540" lazy="no" class=""]

Employees of top 75 Silicon Valley tech firms (race, %)

[visualizer id="13532" lazy="no" class=""]

Most of the problems of ‘diversifying’ stem from internal processes which set up barriers that make it difficult for people to overcome. It is not from a lack of talent, but the fault of the tech companies themselves for not actively and meaningfully engaging in diversity, equity and inclusion.

Namely:

Biases in recruitment pipelines:
Steady stream of workers pumped out of the same colleges and universities ready to enter Silicon Valley

Preference for an elite resume:
Unrealistic achievements and educational requirements for a position

Interviewing process:
Interview questions on subjective traits, or panels that are not representative or diverse

Company culture:
A culture that is not inclusive, or that does not encourage or empower underrepresented employees to speak up and act can deter potential candidates

UNIVERSITY RECRUITMENT.

The first step to understanding the biases in recruitment is looking at where tech companies are recruiting from. One problematic habit is the tendency of Silicon Valley companies to value universities and institutions which have diversity problems themselves. An analysis from 2017 revealed that graduates from schools like UC Berkeley, UC San Diego, Stanford University among others dominate the volume of hires made by the top 25 tech companies. This is in spite of the fact that 

Black students are earning computer science degrees at a higher rate than they are being hired by companies in Silicon Valley

– 9.7% of all bachelor degrees in computer science are awarded to Black students (National Science Foundation, 2014).

At the top of this list was the University of California, Berkeley.

As of 2021, African Americans make up only 3% of the undergraduate students at UC Berkeley and 18% of the undergraduate body come from underrepresented groups (defined by the report as African American, Chicano/Latino, Native Americans/Alaska Native). Besides race and ethnicity statistics, less than 30% of the student undergraduate body are first-generation college students.

If tech companies continuously and actively recruit from the same universities, it is no wonder there is a clear bias in the process. 

A steady stream of workers are pumped out of the same institutions each year. Leslie Miley, the director of engineering at Slack, said that Silicon Valley companies don’t want students with computer science degrees from just anywhere, but degrees from the same schools that the founders and hiring managers went to.

Silicon Valley tech companies thus become “mirrortocracies”, instead of meritocracies. In other words, hiring management and company leaders look for mere reflections of their own backgrounds. Companies have become convinced that only certain schools produce good talent, and oftentimes these schools do not have a diverse pool of talent. These issues of diversity will continue to perpetuate themselves.

INCLUSIVITY, NOT JUST DIVERSITY.

DEI IN PRACTICE

But it is not enough to just hire diversely. Diversity is only effective and sustainable if it exists within an inclusive culture. When people just become numbers to meet targets, this can often lead to a feeling of alienation and neglect which negates the benefits of diversity in the first place.

What needs to be created is an environment where people are able to be recognised for their unique talents and perspectives, and which makes them want to stay. Despite diversity and inclusion initiatives, tech companies still spend more than $16bn annually to replace diverse talent that they failed to retain, due to unfair treatment and a biased work culture. 

spent annually to replace diverse talent after failure of retention

There is a fine line separating empty PR from actual values and strategies – and employees need to know that their company is on the latter side of that line.

Indeed, when organisations claim to care about inclusion, but fail to back up their words with actions, it does more harm to employee perceptions of inclusion than if they had never made claims at all (‘Creating Inclusive Climates in Diverse Organisations’ in Diversity at Work: The Practice of Inclusion, 2014). Therefore, actions which demonstrate the company’s commitment to inclusion are vital.

What practical strategies exist to foster diversity and inclusion in tech companies?

STRATEGIES TO BOOST DIVERSITY.

Companies struggling with homogenous talent pools can invest in reskilling and training programs. Sky did this effectively with a 15-week software development training course for women, which was completed by 215 women between 2016-9 (The voices of our industry: BIMA Tech Inclusion & Diversity Report 2019).

For all companies, ‘blind recruitment’ (removing personal information from CVs) helps to reduce unconscious bias. Applicants with difficult-to-pronounce names fare worse than applicants with identical CVs but easy-to-pronounce names. (What Your Name Means To Your Chances Of Landing A Job, Cole S., 2014)

Finally, granting bonuses to employees who refer diverse talent for jobs also increases diverse applicants. (The future of diversity and inclusion in tech, Dickey, 2019). Of course, strategies to improve diversity must be complemented by inclusion to be sustainable.

STRATEGIES TO BOOST INCLUSION.

Inclusion strategies are as much about commitment as they are about specific techniques. For an inclusion strategy to be more than toothless corporate reputation management, it needs the commitment of leadership at the C-Suite and middle-management level. This is important as ‘diversity fatigue’ and backlash will need to be weathered by those leaders.

One way of instilling this commitment is with a dedicated diversity and inclusion executive reporting directly to the CEO and with power to hold other executives to account.

As well as the backing of leadership, clarity is vital for effective inclusion strategies. Drafting a business-wide strategy on inclusion is recommended by numerous reports (McKinsey, BIMA, PwC, to name a few) so that policy is consistent across the company and employees know their rights and responsibilities regarding inclusion.

But what exactly are these strategies that these leaders should commit to?

Formal training sessions are used by some companies to address issues such as unconscious bias. However, evidence suggests that these alone do not have a statistically significant effect, and can in fact be counterproductive if they prompt backlash or complacency.
(The future of diversity and inclusion in tech, Dickey, 2019)

More effective is integration of inclusion into existing policies – for example, considering how promotions, raises and recruitment will impact employee perceptions of workplace culture. If favouritism and cronyism dictate rewards, then no amount of corporate strategising can create an inclusive culture.

‘Creating a culture’ is difficult partly because it is dependent on individuals, often middle management. Surveying employees to identify which managers are succeeding and which are failing is vital; enabling companies to hold managers to account and share successful strategies.
(‘From Diversity to Inclusion: An Inclusion Equation’, in Diversity at Work: The Practice of Inclusion, 2014)

Ultimately, the individual strategies adopted by these managers can vary. One manager in a university encouraged faculty members to create posters about their hobbies and lives. Despite initial scepticism, it worked: bringing people closer together as they found common ground. This enabled new research collaborations and the smoother running of committees.
(‘Creating Inclusive Climates in Diverse Organisations’ in Diversity at Work: The Practice of Inclusion, 2014)

This is just one example; the takeaway is not to replicate this, but rather to incentivise managers to foster inclusive communities.

In conclusion, values of inclusivity must be guaranteed by leadership but enacted on the ground by managers.

RECOMMENDATIONS.

DIVERSIFY SEARCHES FOR TALENT.

Active recruitment at universities, schools and institutions that predominantly centre minority populationse.g Historically Black Colleges and Universities (HBCUs), Hispanic-Serving Institutions (HSIs), Women’s Colleges

Look beyond traditional degrees/backgrounds: coding camps, community colleges, vocational training programs

Utilise Twitter hashtags: #girlswhocode, #blackgirlscode, #ilooklikeanengineer

Programs for underrepresented groups e.g Black Girls Code, Lesbians who Tech, TransTech Social

CHANGING COMPANY LANGUAGE AND ATTITUDE TOWARDS DIVERSITY.

Show a genuine commitment to representation – provide a diversity statement on your company website; publicly share demographic statistics and goals

Stop using “culture fit” and instead think about “culture add” – what can candidates bring to your team that you may not necessarily have already?

COMMIT TO FOSTERING AN INCLUSIVE ENVIRONMENT AND CULTURE.

Ensure leaders are committed – compulsory diversity and anti-discrimination training and the provision of diversity mission statement where those at the highest management levels are formally committed

Implement code of conduct and non-discrimination policy that is taken seriously and enforced at every level with severe consequences

Create a clear strategy that applies to the entire company – e.g formal training sessions for anti-discrimination and diversity

Gather survey data to see which strategies work and then adapt accordingly

DataEthics4All

JOIN US.

Keep in touch with the latest in Data Ethics, Privacy, Compliance, Governance and Social Corporate Responsibility.