On 16 February 2021, we hosted our popular event Diversity & Inclusion Technology: The future is now. The event aimed to explore how employers can harness technology to build strong and diverse teams faster and better. There were numerous questions from the audience and our hosts Jennifer Millins (employment partner and discrimination expert) and Dan Sinclair (head of MDR LAB) that we did not have time to discuss on the day, so we interviewed one of the speakers Khyati Sundaram, the CEO of Applied, to find out more about her views on how hiring technology can help improve diversity and inclusion.
Jennifer Millins: For a company that does not capture or monitor any diversity and inclusion data, where do you suggest they start?
Khyati Sundaram: Speaking with hiring in mind as that’s our bread and butter, there is a huge variability in HR practices, not least because of lack of data. A lot of companies are at the nascent stages of their hiring and DEI journey and it’ll be natural that they have to start small. For those companies, there are a couple of tactical options:
- Small companies could start in a manual way (think a Google form) and start collecting equal opportunity data, when they are asking for candidates to apply to their job vacancies. This will form a good foundation for analytics later in the process. A caveat - companies need to ensure that they have equal opportunity policies in place and their GDPR compliance is up to scratch before they embark on this journey. To get the ‘right categories’ on equal opportunities, companies need to straddle what’s most important to them and what’s protected by law in terms of demographic data. At Applied we ask for seven different dimensions for diversity when candidates apply for jobs and our disclosure rates are ~90%, so we know that this is achievable.
- The second way and the easier way is to assess how much companies will be hiring in the year and use a tech solution as they scale. Even if they want to add five to ten people in the year to their team, they may find that a tech solution is worth the cost. They will of course need to check that the tech vendor collects this data and removes the pain of GDPR compliance for them. We argue that, being at the start of the journey is in fact an advantage as it’s the best time to start building your hiring systems: before there is an equality and diversity debt in the team, as we often see with scale-ups.
More insights on hiring with data are available in Applied's guide to getting hiring right in start-ups.
Jennifer Millins: What support do you think organisations need to proactively build diverse talent pipelines? Does tech play a role there too?
Khyati Sundaram:
Scrutinising thousands of hiring funnels, we can see that there is a lot we can do to build diverse pipelines, especially in sourcing and evaluating candidates.
Many of the current sourcing systems rely heavily on outbound / LinkedIn; recruiters spend most of their day on LinkedIn these days. Ensuring that recruiters are aware of biases that can creep in is a start, although we know that we can't train individual biases out of the human brain. We need systems to guardrail people’s decision making when it comes to building a diverse candidate pipeline. To be confident that you are attracting all minority groups and not inadvertently perpetuating bias from Linked resumes / photographs, hiring teams need to be intentional about their brand, their entire hiring funnel and even how they write their job descriptions. We know that heavily gendered ones deter women and ethnic minorities from applying. Using our inclusive job ad tool, we’ve been able to increase applications by women to almost 10 percentage points over 300k applications.
Applied is also battling the myth that there is a shortage of talent. We believe the problem lies in the process and not just in the pipeline. So even if organisations manage to attract candidates from minority groups, they often drop out during the hiring process. To maintain the integrity of the process, we advise using de-biased mechanisms to run your hiring. De-biasing the entire funnel is paramount – especially if you use ‘blind’ CVs/ resumes to screen candidates, you’d need to ensure that bias is not being funnelled into later stages like the interview stage.
For more on removing bias from your own processes, see Applied's guide to debias your entire hiring process.
Jennifer Millins: Why is analytics becoming increasingly important for businesses wishing to improve diversity?
Khyati Sundaram: Simple. You are what you measure. Everyone talks about data driven recruitment but most people are not doing it or not doing it right. Analytics is the foundation upon which you can build good hiring practices and help reliably predict what works and what does not for your organisation. There are many ways to approach analytics for hiring but starting at the basics of hiring metrics may help build the right analytics level for your company.
You can read Applied's guide to hiring metrics and reporting here to assess your data analytics needs.
Jennifer Millins: What are the major obstacles preventing fair pay within companies and how can companies start to understand and tackle their gender pay gap?
Khyati Sundaram: There is much research showing pay decisions are influenced by many of the same kinds of biases we see in hiring. At Applied, we focus on the hiring part of the employee life cycle for now as it is often the first and most important place to make change. That said, the Applied product does contain some small nudges to try to support better decisions around pay. This includes things such as nudging hiring teams to make salaries public on job descriptions and to note if they are negotiable, both of which have been shown to significantly reduce gender pay gaps.
Jennifer Millins: If you are collating data for analytics purposes, who do you share it with? How do you protect it?
Khyati Sundaram: There are multitudes of legalities to consider here. You should first assess what you will be using the data for, build your use cases and then your privacy policy / T&Cs to incorporate those cases. Then comes the not so tiny matter of GDPR compliance. GDPR compliance is tricky with the best of intentions, so I would advise getting a good lawyer or compliance expert. Larger companies are required to have a data protection officer (DPO) when dealing with sensitive data. But when we are dealing with data such as candidates' data, I would recommend getting a DPO to ‘stay’ on top of your compliance needs and frequently keep addressing any gaps. Personally identifiable data (PII), according to GDPR cannot be stored forever, so you could always delete personally identifiable data as soon as the hiring process is done and/or anonymise all data and run analytics on that in the long term.
For advice on data protection and GDPR compliance, contact Mishcon de Reya's Data team.
Jennifer Millins: How does your tech stack play into the overall broader diversity strategy? Are organisations relying on you and your businesses, or are you a piece of the broader puzzle?
Khyati Sundaram: Speaking for Applied, we are a solution for de-biased hiring. We solve for predictive assessments that will ensure you select the best candidate every time. We also solve for a small piece of sourcing aka inbound sourcing. So, organisations are relying on us for hiring diversely. But hiring correctly and diversely is only one part of the puzzle. There are other pieces to the puzzle such as retention and promotions and making sure the organisation is being inclusive on all aspects of employee lifecycle. While we at Applied don’t solve for the other pieces, there are incredible tech companies out there who can help with your D&I strategy such as GapSquare and FairHQ.
Jennifer Millins: As with any technology, selling to a large organisation is hard. Is it easier when selling something that is so top of mind of leadership?
Khyati Sundaram: Not necessarily. Change is always hard! We know none of the incremental solutions towards D&I have worked in decades and so, new revolutionary solutions need to be created. These will be hard to sell into the organisation if the organisation has had long-standing ingrained systems. But we believe with a structured process, selling this change (and buying into change and implementing that change) can be made easier, especially if that structure includes empowering actions towards the goal.
Applied's recruitment change management guide looks at change management frameworks in the context of overhauling hiring systems.
Jennifer Millins: What should organisations be doing right now to ensure their D&I efforts are sustainable long-term?
Khyati Sundaram: Organisations need to identify structural interventions that can cultivate lasting change. And that's a slow burn - there is no short-cut to diversity. As a research based organisation, we like to use science where we can to help us navigate. A lot of the evidence says that we can't train bias out, we can't solve for this problem in the way we have been solving it for decades. For example, we can't employ quotas and expect that to be sustainable long term in organisations as it creates considerable backlash particularly from groups who feel threatened by them. And that backlash can become a hindrance to adoption. If we can’t use old processes, how do we start thinking of new architectures that can bring about sustainable change? Sustainability will come from intention, constant iteration and a culture renovation - all the things that take time!
Jennifer Millins: What trends do you expect to see from the D&I tech industry over the next few years?
Khyati Sundaram: I am opening Pandora's box here but a big impending trend is AI. It's the red hot sizzling topic technology that is being used in candidate matching and hiring assessments. We should recognise that AI is a nascent technology and great care needs to be taken when it is used in people decisions. Algorithmic bias and biased training data sets can lead to perverse or outright prejudiced decisions. We are cautiously optimistic though – we’re experimenting with uses to identify 'safe' ways of building it ourselves. The crux of the problem is that the vast majority of AI in recruitment tech is focused purely on efficiency – speeding up the process to for the benefit of the hiring organisation. Applied is on a path to find sustainable implementation of AI as we want to find what is best for both candidates and organisations to realise the full benefits of AI in hiring.
About Applied
Applied is a hiring platform that uses behavioural science to remove unconscious bias and improve predictive validity. They have transformed hiring by using research-backed methods to replace traditional methods of hiring (like CV/resume screens) with more predictive assessments and designing out bias from candidate scoring by anonymising candidate answers. Applied is the first tech spin out of the UK’s Behavioural Insights Team and was developed in partnership with Harvard Professor, Iris Bohnet.