Insights ›
Using Data to Improve Your Diversity and Inclusion Strategy

Diversity and Inclusion is a hot topic in the world of talent acquisition, but how can we optimise our D&I strategy with a data-driven approach?
Niamh Joyce explores how organisations can double down on their D&I commitments by analysing the data behind their approach.
Diversity and inclusion is a true ‘hot topic’ in all areas of society, not least in the workplace. The average person will spend 90,000 hours at work over the course of their lifetime – time enough to watch FRIENDS more than 1,000 times. That’s a lot of time to spend feeling unseen or underrepresented; the value of a diverse workforce is much discussed in the world of HR, and more often than not, a large chunk of the responsibility for ensuring diversity falls to Talent Acquisition.
At a recent session of the Elements Data Guild we explored Algorithmic Bias (with the help of Joy Biolawimi’s 2017 TED Talk) which led into a conversation on the importance of Diverse hiring, and while we all agreed on the importance of being aware of this matter, there’s a fine line between active inclusion and box-ticking. When we talk about diversity at Elements, we include attributes such as gender, ethnicity, sexual orientation, religious beliefs, physical abilities and class background within the term. Hiring for diversity is not simply about hiring with the goal of looking good on paper, or being able to spout superficial statistics about the split of your workforce. Instead, we view ‘unbiased hiring’ as a scientifically inspired approach which aims to minimise the influence of hirer bias, allowing us to assess applicants meritocratically. The only way to truly recognise, and eliminate, bias is to ensure that our Data matches our needs. Capturing the right information, and doing so accurately, will help to ensure that your D&I strategy is fit for your organisation, and to pick out pain points before they bubble to the surface on their own – as these things usually do.
At Elements, we’re lucky to have a custom-built data capture strategy which helps us in taking this scientific approach to fair hiring. However, most of the data points discussed are easy to capture, either manually or within an ATS.
Understanding your organisation and setting goals
‘Diverse hiring’ is not one size fits all, and it’s vital to start with goals that are practical, relevant and fair. This means understanding what your organisation looks like and, where possible, why. It’s one thing to be able to say that you have a high diversity ratio overall or a 50/50 gender split. But how is this weighted at each level? What does diversity look like at the decision-maker level for your individual teams? Who is in a position of leadership, and who’s being promoted? And again, this is not just about ticking boxes – think about the impact this is having on your user/customer experience or service offerings. Without capturing at least some of this information and getting a true understanding of your organisational structure and map, it’s impossible to create true goals for diverse and unbiased hiring.
Workplace surveys are the best way to get this information – it’s possible (and simple) to anonymise these enough to protect sensitive information, but maintain the ability to break these down by teams/levels.
Attracting and accommodating to all the right candidates
Coded language in job adverts is not a new conversation in the TA world. Many hiring professionals will have heard various statistics about candidates being more or less likely to apply for a role containing particular descriptors, and a multitude of tools (Textio, gender-decoder) exist to help avoid unconscious bias in job adverts. These tools, which often aim to eliminate implicit bias by identifying tone and so-called ‘gendered wording’ can be a good starting point. However, there’s a risk that seeing that green tick will lull you into a false sense of security – not to mention their basis in the assumption that women applicants still find words like ‘ambitious’ to be masculine. Instead, collect information about the diversity of your pipeline at the top of the funnel. With applicants, using tools like LinkedIn’s pipeline builder will let you know who is looking at the job advert, or starting the application, but not completing it. Where are you posting the job? With sourcing – how diverse is your sourcing pipeline? And who is actually responding?
Coded language is one barrier, but so is not advertising in the right places, too much specificity – even including specific educational requirements – could isolate multiple large groups of otherwise qualified candidates.

As this visual demonstrates, the STEM world’s issue with diversity is present even at the level of higher education. By including hard educational requirements like Bachelor’s or even masters degrees, which is very common in the tech world, just how many capable hands are you excluding? Especially taking into account what we know about certain groups of applicants being less likely to apply if they don’t meet these requirements.
Capturing the high-level data may not be enough to tell you what your specific issue is, but it will be enough to tell you that the issue is there, and the earlier you’re able to identify a challenge, the easier it is to solve.
Understand Advertising
On the flip side, it’s really important to understand where your applicants ARE coming from. Implementing the ability to track exact source is another key way to help you up your sourcing and outreach strategy, especially when you really get into the detail – this may involve some manual collection when speaking to candidates, but it’s important to understand how applicants are getting to your careers page. If candidates are specifically looking for your careers page because they are interested in joining your organisation, then this speaks to the efficacy of your employer branding. By the same token, it’s useful to know whether people are clicking on a job advert they see on their LinkedIn feed, or a post put out on a forum by an employee, or even Linkedin through an article. This is the only tangible way to understand what channels are impacting what applicants, and to ensure that your presence is strong enough within those channels, to attract the range of candidates that are going to help make your organisation a success. Note – employer branding isn’t just about who applies, it’s also about how much awareness people have of your company identity when you are reaching out.
Track your Process
Once you start paying attention to who is entering the process and where, it’s a much simpler task to keep looking at who is exiting the process, where and why. It’s easy to sigh and grumble over a rejected offer and chalk it up to something like a tough salary negotiation, but part of an effective and inclusive hiring process is taking the time to map out candidate experience. It’s equally possible (and likely, especially if you’re seeing a lot of rejections from one or more specific demographics) that something along the way is giving a sub-par impression of the company or the team. Salary/working conditions are not promoting inclusion, benefits packages are lacking, an absence of flexibility singles out working parents, etc. This means tracking time to hire, reasons for rejection, as well as how these reasons change at various stages. If you collect this information in enough detail then you will be in a position to cross-reference rejection reasons with rejection stages, and get a specific understanding of where the issue lies; if everyone is rejecting at reach-out stage, then it’s an attraction issue. If everyone who rejects after the first interview is doing so because of the role level, perhaps the interview panel or not describing the job accurately, or the questions being asked don’t match up to the expectations of the job.
Track your Performance
This also applies to candidates who are being rejected. Again, assuming that you have a feedback loop in place between interviewers and the recruitment/HR team, this should be fairly simple to track, and is a great way to identify pain points as well as unconscious bias in your process/interviewers. Sticking with the education example from earlier – are all of your candidates who don’t hold a degree falling short at the first stage? Or being rejected by a particular interviewer? Unbiased hiring isn’t just about diversity, it’s about taking everything into account and ensuring that the unconscious bias of one or more interviewers doesn’t overshadow the process of hiring. It’s not a negative reflection on your colleagues, but if everyone on your interview panel is a university educated male from a similar socioeconomic background, it’s likely that (without the right training) they will struggle to see the value of a different background. Equally that your interview process is unintentionally stacked against ESL speakers, who otherwise would be able to bring a lot of value to the team.
Hiring is all about balance – balance of opinions, balance of experience and, of course, balance of diversity. Talent Acquisition is how you set a team up for success – it doesn’t matter how strong your offering, idea or business model is, if you don’t have the right people to execute, then failure is inevitable. Unbiased hiring is not only important at times when society is under a microscope, it is important always, and a robust D&I recruitment strategy must be a conscious effort based on tangible information.