Diversity in data
Hannah Khwaja, Data analyst, Data Orchard
At Data Orchard, we are committed to ensuring that everyone we engage with is valued, supported, and treated with respect.
This means not only making sure that we meet the requirements of relevant legislation associated with equality, diversity and inclusion (EDI), but that we do whatever is necessary to provide genuine equity.
We know that the digital, data and technology fields are still somewhat behind the curve in terms of diversity compared with other areas of work. And that the gaps are particularly stark at leadership level.
As such, I was excited by the opportunity to attend an excellent week of discussions hosted by the BCS Data Management specialist group, BCSWomen and DAMA UK around the topic of diversity (or lack of it) in data (Diversity in Data: Breaking the single source of truth) earlier this year. The talks were packed with brilliant insights - here are some of the key takeaways that have stayed with me since...
No diversity without inclusion
Many of the speakers felt that when employers talked about ‘diversity’, they often really meant ‘visible representation’, with a focus on exterior appearance. But while diversity can technically be achieved by having people in the room, inclusion is about the experience that those people have while they are there.
In an inclusive environment, people will want to stay. And they will tell others about how good it is, which will inevitably lead to further improvements.
Conversely, the opposite is true of a hostile environment, and people are likely to share their bad experience. Hence, there is a serious risk of causing harm by focusing on diversity but not inclusion.
Creating an inclusive environment requires:
Awareness – the ability to perceive barriers
Resources – investment into removing barriers
Commitment – ongoing, centralised efforts to make progress, with leaders that actively demonstrate EDI behaviours and set clear expectations, roles and accountability
Inclusion should itself be driven by data – using insight to assess impact and drive continuous improvement. Numerical data can be used to identify weak points, which can then be followed up with qualitative data on lived experiences. Combined with evidence-based industry standards, this data can be used to inform effective practices.
Embracing neurodiversity
As one of the speakers theorised, data careers are likely to naturally attract neurodiverse people, and ways of working are often matched to neurodivergent thinking.
However, modern data professions tend to ask for a lot of divergent abilities – for example, attention to detail, as well as big-picture thinking; independence, as well as team working; and general business sense, as well as technical subject knowledge. Neurodiverse staff will often have exceptional skills and brilliance in some areas, but may need accommodations to support them in others. Possible examples included someone who is autistic, who may be an excellent communicator, but may feel exhausted after giving a speech or attending a conference and require time to recover; or someone with ADHD who may hyperfocus to meet a deadline under high pressure, but similarly need to be given space afterwards to decompress. Of course, even under the umbrella of these two conditions, an individual’s experience will vary. It’s crucial for employers to recognise this, and for them to build teams that deliver excellence as a collective, balancing each other’s strengths and weaknesses.
Redesigning recruitment
Job adverts and descriptions were identified as being one of the main reasons why data does not have a more diverse workforce – but there were many opportunities identified to improve the recruitment process.
On the positive side, organisations today are becoming more open to having diversity of thought, experience and background. Historically, opportunities in digital and data were characterised by requiring highly technical people in every single role. Increasingly though, organisations see the value of roles such as data ‘translators’, who may be less technical, but more able to communicate complex insights to internal and external stakeholders.
But, let’s face it, adverts for jobs in data are often extremely dry, and simply a repeat of the job description. Yet, data folk are typically curious people. We want learning, impact and progression – and, like everyone, we want to have purpose. The recruitment process should make it clear how people’s work will be visible, valuable, and have an impact – both at the organisational level (the organisation having a values-driven mission) and at a personal level (work making a clear difference within the organisation).
Beyond this, it’s really important that the recruitment process advertises for and assesses skills and experience that are actually needed for the job, without assumptions about requirements (which may be rooted in cultural bias). Some of the issues around this are exacerbated by the largely unstandardised nature of job roles in data. The sector abounds with different titles for the same job, and the same title for different jobs. This can cause problems if an employer isn’t completely sure what they will need from a data role, is constrained by strict HR processes and systems, or if recruitment is led by a hiring manager who is disconnected from the actual position being advertised.
In particular, there are all sorts of routes into a data career, with some taking a wide-ranging path through other types of role. Even for those who have always worked in digital and data, the field changes rapidly. Within the span of one person’s career we have gone from early desktop PCs, dial-up and web 1.0, to today - with smartphones, 5G, rapidly developing AI, and massive collection of personal data.
It’s important that employers use skills-based recruitment to recognise varied routes into data and acknowledge transferable skills, as opposed to strict experience or qualification requirements that don’t have explicit reasoning. Additionally, the influence of unconscious biases can be minimised by masking demographic features and adjusting the interview panel and processes. People are instinctively likely to prefer people who are like them, so a diverse panel is key.
Investing in early-career talent
In order to address the lack of diversity in data, it’s particularly important to take an interest in those who are early in their career – or perhaps not even in the data job market yet. The number of data-focused roles and need for data skills in a wide range of roles are increasing, but almost half of UK organisations say they struggle to hire data talent. We need to welcome new people into data careers, who will bring fresh perspectives and ideas and help to break down stereotypes and barriers.
As data folk, we can support this by offering our time as mentors and advocates, engaging with and offering skills training or ‘bootcamps’ to communities that are currently under-represented, and supporting people as they begin their new careers. A great example is the work done by Grayce, who hire, train and deploy graduate data consultants.
And finally…
It’s important to note that initiatives that promote diversity, equity and inclusion make good business sense, and they will benefit everyone – even those who may not feel that EDI is relevant to them. We all have times in our lives when flexibility in policies, practices and behaviours will benefit us, for example if we have children, need to take on other caregiving responsibilities, or suffer from periods of illness.
EDI is a cyclical process. Continuous monitoring, evaluation and informed action is essential. We need to maintain our passion for people, as well as our passion for data.
The work is never done.
Thank you so much to Tristi Tanaka, Prerna Tambay, Jenny Andrew, Marco Bernardi, Grace Treacher, Tosin Sonubi, Paul Dettman, Roisin Wherry, Catherine King, Kyle Winterbottom and Lisa Allen for their insights, and to the BCS Data team, BCS Women and DAMA UK for hosting this event.