Follow Quality

Algorithms: the solution to recruitment woes?

In Quality, Talent, Teams - 3 years ago - 5 min

Algorithms: the solution to recruitment woes?

When it comes to recruitment, algorithms can play a role, but there are still times when a personal touch is needed.

If there’s one common complaint to be found among entrepreneurs, it is that it’s hard to find good staff. Business leaders can spend hours scanning LinkedIn, attending networking events, or shelling out cash to agencies and head-hunters. That’s where data-led, algorithmic recruitment can help. Using an algorithm isn’t only a quicker and cheaper way to find better matches, it can also lead to higher retention rates and it reduces potential bias too.

Currently, algorithms are most often used when hiring in the IT sector or by large companies with lots of data, however you can expect new applications and services to crop up soon to help you with your recruitment woes. But, by placing so much faith in number crunching, are we in danger of losing the human touch?

Entrepreneurs Hung Lee and Gordon Dent felt they were regularly being contacted for “jobs they didn’t want to do anymore”, and believed a more effective solution was possible. So they built a platform,, where coders and developers specify the kind of job they are looking for and the algorithm finds the best match. Lee says algorithms shouldn’t be looked upon with suspicion as they are “simply sets of rules” which, in this case, match preferences with offers. He also says this algorithm-based approach is a more effective form of recruitment than CV orientated hiring. “Everyone has the historical data – the CV – but no one has the future data – what this person wants to do next,” says Lee. “We have seen employers go from sign up to interview a live candidate in less than 10 minutes. Typically, companies hire on the platform within a week of signing up to it, while the average time to hire a software engineer elsewhere is 90 days.”

Lee suggests that the CV has had its day, although convention might mean they will last longer than they should. “CVs have been part of the recruiting landscape for hundreds of years, many parts of our recruiting practices remain entirely dependent on CVs. But are they the most fit for purpose format of data that we have to move information about people around the internet? The answer is clearly ‘no’,” he argues.

Furthermore, Lee says it’s what people want, and that the numbers justify this view. “I have collected over 2,000 unsolicited email messages written to me personally, from developers all over the world, urging us to continue our work. We also have tremendous feedback from clients.”

Supply and demand

The crucial part of any algorithm is having good data, and having lots of it - algorithms only really work at scale.

Job matching algorithms are being used to create new kinds of businesses. One of them is Tutora, set up by Mark Hughes and Scott Woodley, which is designed to match students with private tutors. The business, which began in August 2015, has over 2,000 tutors and collects a commission for each lesson taught. Hughes says algorithms are effective but require lots of data, so sites such as his require a critical mass. “The crucial part of any algorithm is having good data, and having lots of it – algorithms only really work at scale. We record every enquiry made on the website, and whether that enquiry was filled or not. We then group these together by subject area and location, and if our “fill rate” falls below a certain threshold, we know we need more quality tutors in that city,” he says.

But the human element cannot be ignored, and Hughes says the business has to ensure there is a balance between demand and supply. If there are too many tutors, there won’t be enough work. If there are too few then customers will become unsatisfied. But he says that automated advertising tools are part of the answer.

“Our main challenge is to match demand with supply. So, if we know there are lots of requests for SATs tutors in Birmingham, but we currently don’t have many tutors there, we can advertise for tutors in those specific locations for those subjects. The new wave of online recruitment channels such as Indeed or Google AdWords means we can be really targeted in our approach, turning adverts on and off as we need them,” he says.

Algorithms do appear to be a disruptive force in the recruitment sector and leaders in the sector recognise it. David Spencer Percival, CEO of recruitment business Spencer Ogden, still relies on more traditional methods such as CV, interviews, assessments and work trials. “Despite the benefits technology brings to recruitment, you can’t write an algorithm as a replacement for two humans to interact in a challenging conversation. At least not just yet, anyway,” he says.

Percival says the human touch is still required, as employers want people who can work together and communicate. He suggests that such criteria are subjective, rather than rule-based. “Our candidates are technical, but not robots. They need to interact with other engineers and project managers, and work as a team to complete complex tasks such as deep water drilling, or building a nuclear power station. Personality and communication skills, as well as technical skills, are required.”

However, Percival recognises that change has come to the industry, and he is considering new, data-based forms of selection to complement existing techniques. “I’m a huge fan of technology and believe a good algorithm could match skills, ask interview questions, map trends and possibly even habits and traits effectively. However, and as with all technology, this depends a lot on the data and the input of available information,” he says.

Algorithms may not be widely used – yet. But looking at the retention data, it’s worth asking your recruitment provider if they can use an algorithm the next time you’re searching for the right candidate.

If you are looking for support to grow your business please get in contact with one of our team on 08081 722350 or drop us an email: