Headstart CEO Gareth Jones: Enabling Diverse Workforce with Machine Learning

How Data Science warrants a robust match between a job and a candidate, levelling the field for all fitting candidates.

Yuliya Sychikova
COO @ DataRoot Labs
01 Apr 2020
6 min read
Headstart CEO Gareth Jones: Enabling Diverse Workforce with Machine Learning

Gareth Jones is the CEO of Headstart, which transforms the way organisations hire and enables clients to hire a more diverse workforce. The company uses Machine Learning and Data Science to transform the recruitment process, enabling clients to find the right high potential employees regardless of gender, ethnic status, sexual orientation or age.

Founded in 2017 in London, Headstart has raised $7 million in seed funding round led by FoundersX with participation from Founders Factory.

How was the idea of Headstart born? Why have you decided to head the company as its CEO?

The idea for the company came to one of our founders, Nick Shekerdemian, who had dropped out of Uni and started an online business in China, where he had originally been studying. In conversation with his friends who were finishing Uni, they pretty much all complained about how clunky, out of date and especially unfair the graduate job application process was and so the idea of Headstart was born. The other co-founder, Jeremy Hindle, the CTO, had previously built and sold a couple of gaming companies but had also created the largest minecraft server for people with learning difficulties. So fairness and levelling the playing field has always been in our DNA. Originally conceived as a marketplace for graduates and employers, we pivoted to a full SaaS solution in the summer of 2017 when we went through the Y-Combinator program in the US.

Having joined the company as an advisor in January of 2017, I moved full time as the COO in September of that year as we exited the YC program. The decision to take on the CEO role was the result of a revaluation of the core roles and capabilities across the team. It is very exciting to lead such a great business in a field - HRTech - that I have been in my whole career.

Headstart aims not only to reduce time-to-hire but also to make the hiring process less biased and more efficient. How the data driven technology helps you accomplish that?

We focus on what matters and not what doesn’t. We use various data sources to rank candidates as they apply, and match them properly with the requirements of the job. The vast majority of recruiting is done using a simple (and usually out of date) job description to define a job and a CV to define the individual, with either a basic match done on skills and experience by an ATS or a human being.

The simple truth is that if you want to hire accurately you have to scientifically measure the job and the candidate in order to find a proper robust match. It isn't rocket science, but the old method prevails to this day. At Headstart, we use a combination of data sources including company, existing employee, market data and candidate data in order to create an accurate job profile or fingerprint as we like to call it and to measure the candidate’s fit to the job.

By increasing accuracy and focussing on the right data points, recruiters can focus their energy on those candidates who are a proper fit i.e those with the highest match scores. This not only saves them time by not having to scan through thousands of CV’s but it also presents them with a more diverse candidate pool, as they are not doing initial screening and falling into the trap of letting their unconscious biases influence those initial screening decisions. Candidates, screened and measured on the right attributes, are presented in terms of fit for the role based on relevant data points and not excluded because of their ethnic origin, sexual orientation, age or gender.

AI is as good as its data. Headstart claims to achieve15% bias reduction for Asian ethnic minorities, 18% bias reduction for Black ethnic minorities. How do you ensure that your data is bias free and your algorithms do not exacerbate the problem but help solving it?

We measure. At every step. Our system is able to track diversity data through the entire recruiting process and we can highlight where approaches may need to change. One of the challenges is that most companies have no idea how biased their recruitment process is. They only have the current company makeup as a yardstick, which is after the fact and in most cases, lacking in diversity.

Most companies don’t realise that their processes and the technology they use exacerbate the issue. It is actually very difficult to analyse the process. We were fortunate enough to have to do a parallel test with an existing process and technology stack and that A/B test showed just how disadvantageous the current process was to Asian and Black candidates.

Your algorithms are designed to find the “best fit” talent. However, “best fit” is quite a subjective term. How do you go about it? What parameters do you tune your results on? Is any of the process manual?

Best fit is indeed subjective when not measured properly. And 90% of companies don’t measure it properly, if at all. Most companies base recruitment fit on a candidate's experience i.e. have they done this job before in our sector or company size. Very subjective! It actually doesn't take much effort to go beyond this and look at skills, capability, behaviour and other traits. And if you measure the job properly (and don't rely on the job description) then you are significantly reducing subjectivity.

As I say, whilst the notion of accuracy in hiring is often discussed in a way that frames it as a holy grail, it really shouldn’t be. It really isn’t that difficult to make hiring more accurate. Generally speaking, the problem is that companies are simply just too lazy or just don’t value the people in their business enough to do what is required.

I find your list of customers very impressive - Accenture, Lazard, Smiths among others. Is it fair to say that larger enterprises are more willing and ready to improve their hiring process? What challenges do you face while acquiring new customers?

I don’t think the desire to improve hiring accuracy is better in larger companies. I would even go as far as saying smaller companies care about it more, simply because a bad hire in a small company has potentially more fallout. But they still tolerate it nonetheless. Companies that see the value in their people and understand people science are the ones who take hiring accuracy seriously.

When it comes to acquiring new customers, one of the biggest challenges is undoing the complexity that has been built into their process, usually to accommodate the poor enterprise technology that supports it. That and poor processes that have been built on and made more complex over the years, seemingly become folklore and or are seen as essentials when in fact they are not. The biggest barrier to great hiring is complexity. We live in a world where technology is ubiquitous and as such, we have forgotten how to re engineer processes properly before we adopt technology. We continuously build on complexity when we should be striving for simplicity.

I don’t have to tell you that one of the most common discussions around recruitment is the ‘experience’, particularly for the candidate (and the recruiter). We continually strive to improve the overall experience and make it better so we convert more candidates and create a better impression as an employer etc etc yet we build recruitment processes that are way too cumbersome. Instead of simplifying, all the technology has done in the last 20 years, especially in large scale enterprise recruiting technology, is ‘automate chaos’ as I call it.

The biggest barrier to great hiring is complexity. We live in a world where technology is ubiquitous and as such, we have forgotten how to re engineer processes properly before we adopt technology. We continuously build on complexity when we should be striving for simplicity.
Gareth Jones
CEO @ Headstart

Congrats on the $7M recent fundraising round closed in November 2019! What features are next for your clients? What do you hope to achieve 3-5 years from now as an organization?

Thanks to the investment we can now double down on product development. That starts with product discovery - identifying the problem to be solved. A lot of product companies, many in our sector too, jump straight to a list of features that either came direct as requests from customers or from the sales and marketing teams. This is old school product management. Customer inputs are valid, as are inputs from the sales team. But a list of wants doesn’t necessarily translate well into what is really needed, what the job to be done really is. And it doesn’t translate into a great product that solves that problem well.

So we don’t talk about features in that context. We talk about what problem we are trying to solve. We research that with clients, then work that through a process that eventually delivers value in the product, through improved usability, simplicity etc.

In 3-5 years we want to be category leader, in entry level hiring. We want to be responsible for delivering the leaders of the future into enterprise companies. A more diverse set of leaders who will then in turn create more diverse organisations.

In 3-5 years we want to be category leader, in entry level hiring. We want to be responsible for delivering the leaders of the future into enterprise companies. A more diverse set of leaders who will then in turn create more diverse organisations.
Gareth Jones
CEO @ Headstart

What do you consider your biggest accomplishment as the CEO?

I think the biggest and most important role for any CEO is to get the very best out of the people in the business. My role is to remove barriers that are getting in the way of my teams doing their job, then get out of the way. Letting go of control, pushing down responsibility and giving the people in the business the autonomy and latitude to deliver on their outcomes is the main objective. But it’s hard. Traditionally we have seen the CEO, and the construct of the traditional organisation to be one of control.

Trusting your people to get on and deliver on the brief you gave them, without meddling or interfering is something that is hard for a lot of business leaders as they were schooled to be in control, in charge, make the decisions, call the shots and monitor what people are doing in the company. I have no interest in that. I have hired a great team for one reason - they are (hopefully!) smarter than me at what they do, and will do a great job, as long as I let them do it.

My job is to set the magnetic north, make sure our people are in the right jobs, happy and healthy and make sure we always have enough money to continue our journey. In our business, the individual comes first, not the company. It is very important to me that everyone in the business can be and is their “best self”. And we support that as a fundamental to the health of the business. If you are your best self, then you will deliver your best self to your team, and in turn that team of ‘best selves’ will deliver the best to the company.

Our mantra is very simple. We don’t care where you work, when you work, or how you work. Just meet your outcomes and commitments to yourself first, then your team and then the business.

What was the biggest mistake you’ve made as the CEO of an AI startup?

Not being transparent enough. There is always the temptation (largely because it's the traditional view of how to run a business) that you should not tell the whole story to the team, especially in difficult times. Not share the real difficulties, the biggest derailers or the gloomiest of outlooks. The thought is to share just enough. Keep something back either because they don't ‘need to know’ or because they may ask difficult questions for which we may not have any robust answers. Every time I have fallen into this trap, it has backfired.

If you want to build trust, which is a fundamental building block of a great organization and employee engagement, then you have to be honest and open. This isn’t about anything more than treating people like adults. For some reason, the moment an employee walks in the door, we suddenly stop treating them like adults, confining their job responsibilities, limiting their budgeting sign off, putting constraints around their decision making scope. And the worst, withholding information. The more open you are, the more trust you build, the better the business.

Currently Headstart employs over 15 people across the UK, US and other countries. How does your product impact the culture within your organization and the way you hire?

As a small business we don’t actually use our own product, our hiring volume is too low and doesn’t generate enough data. We do however, practice what we preach so we manually measure each role scientifically and create a profile, against which we then measure each candidate. This means that each candidate goes through a process of assessment that matches them to the assessment we did of the job role itself (which is also done using the same assessment tool criteria).

This means we focus on values, motivations and behaviours and we have a clear understanding of what is required in the role and also how the candidate fits. We share these results with everyone in the company and they become a common language around which we understand each other as human beings.

Now, as you very well know, there is a certain unease among people about AI in general as AI is expected to replace jobs. Do you think Headstart is contributing to this notion? Which parts of the recruitment process does Headstart fully automate?

I think there is a major misunderstanding of what AI actually is and a completely distorted view of how it is used and how much it is used, in this industry in particular. The problem is the hype around it has exploded beyond imagination and there is so much total rubbish talked about that it’s hard to get to the real facts.

Firstly, there is very little real ‘AI’ in recruitment. Recruitment, as i said, is not rocket science! And it doesnt need to be turned into rocket science either! What we are talking about is data science and some machine learning at best. Clearly they both come under the banner of AI, but using the term AI so widely and glibly serves no purpose at all apart from to confuse or spend marketing dollars.

In most cases, just like us, we are using data science to look for patterns or correlations in data sets. Signals. Indicators. Things that we have identified as being valid, relevant and key for a successful match/hire to take place. We use these outcomes and build them into algorithms (which have been around for centuries!) in order to automate the process of screening and sifting, NOT hiring. The headlines have become obsessed with it and have created a largely nonsense dialogue around it.

In terms of automation, well yes of course we are. Any technology is by default, by its very existence, automating something. Automating a previously manual or semi manual process. But this has been going on for years. And the extent to which it can reach sci fi levels of automation in recruitment particularly, is limited. It just isn't that complicated a business process. Headstart has an impact in two ways. Firstly we automate the screening process, so we remove a large proportion of the time a recruiter would have spent simply going through CV’s and trying to sift out those that could be a match. As it stands, a human being is not capable of understanding the core matching criteria (scientifically measured) holding these criteria in their head, across several positions, and accurately scanning a cv to determine an objective match to the job. It’s just not possible.

Humans are subjective in these conditions. So in our context, we automate a subjective and manual process, increasing accuracy and efficiency. And releasing that recruiter to do more value adding and frankly interesting things in their job. Can it result in fewer recruiters needed? Yes of course. But the reality is it will not shrink the global volume of recruiters by 95%, which is what you would be forgiven for thinking if you listened to the hysteria.

What do you think is next for HR tech? What in your opinion will we see in the next 5 years? Who will be the new market winners and losers?

HR Tech has always been the poor cousin of enterprise technology, and still is, despite the surge in interest as a category for investors and entrepreneurs. The tech and techniques are still light years behind what's available in other functions within business. In fact, many of the technologies and approaches we are hailing as ‘new’ and ‘innovative’ in people tech have been alive and well in other parts of the organisation for nearly a decade.

The upside of that is that it has plenty of runway yet. I think it’s a good bet for investment and has plenty of room for innovation. As companies eventually start to get with the program and realise that people are their only point of competitive advantage, they will start to invest properly in people, and the tech that surrounds them. One day, a long time in the future, but at some point the tech we have at our fingertips that we use to manage the people in our business will give as much insight into our people as our existing customer tech gives us into our customers. At the moment, the gap is enormous. We are only just scratching the surface.

The other trend you will see is the consumerization of B2B technology. HRTech especially is clunky, old and isn't designed for the most important people in the mix - the users. As consumers we now have access to tech that supports our personal (and professional) lives that we pay for and use ourselves. We have become very discerning in this context too. We expect excellence. Tech that has 10 killer features, not 1000 useless ones, is beautiful to use, can be used on any device and fits seamlessly into our lives and, most importantly, drives value for us as individuals, wins the day. If it doesn’t, it dies. Enterprise HRTech of the future will have to meet the same criteria.

Have an idea? Let's discuss!

Book a meeting
Yuliya Sychikova
Yuliya Sychikova
COO @ DataRoot Labs
Do you have questions related to your AI-Powered project?

Talk to Yuliya. She will make sure that all is covered. Don't waste time on googling - get all answers from relevant expert in under one hour.
OR
Send us a note
Optional
File requirements pdf, docx, pptx

Author

Yuliya Sychikova
COO @ DataRoot Labs
Yuliya is a co-founder and COO of DataRoot Labs, where she oversees operations, sales, communication, and Startup Venture Services. She brings onboard business and venture capital experience that she gained at a leading tech investment company in CEE, where she oversaw numerous deals and managed a portfolio across various tech niches including AI and IT service companies.

Co-Authors

Ivan Didur
CTO @ DataRoot Labs
dataroot labs logo
Copyright © 2016-2024 DataRoot Labs, Inc.