Predictive analytics and algorithms: the future of recruiting?
My inaugural piece for the HR Tech World blog provided 10 tips to getting started with HR Analytics. Now it’s time to get more practical and turn our attention to recruiting, arguably the biggest showcase for a data driven and analytics led approach in the HR sphere.
Recruiting is the perfect shop window for analytics as not only is hiring great people of utmost importance for all organisations, but hiring is high-volume, comprised of repeatable processes and for too long has been primarily based on intuition and unconscious bias.
The cost of getting it wrong is significant, with a recent report in the Telegraph estimating that the average cost of replacing a departing member of staff being over £30,000. This is actually more conservative than other studies by the likes of PwC and CEB that have found that the replacement cost is actually one and a half to two times salary cost. However, with the market for top talent being highly competitive, the pressure on recruiters to hire quickly coupled with an almost fundamentalist emphasis on minimising cost per hire, quality is too often sacrificed.
This lethal combination of gut instinct – surely tantamount to guesswork, and the focus on metrics such as time to hire and cost per hire over quality of hire is potentially dangerous. Getting it wrong isn’t only costly, poor hiring can lead to lower productivity, reduced levels of employee morale and engagement and ultimately more attrition. It is a vicious circle.
So what are the alternatives? Well, firstly being less obsessive about the cost per hire metric – use it as one of a number of metrics to guide the effectiveness of your recruiting program not as the single source of truth. Indeed, as I wrote recently (see ‘In recruiting, how important is Cost per Hire?’) – utilising insights from a Bersin by Deloitte study, placing too much weight on cost per hire is not only counterproductive but also more expensive in the long run. In summary, cost per hire is a false idol for recruiting functions and should come with a big health risk.
Secondly, analytics – specifically predictive analytics – should also be deployed to reduce reliance on the gut instinct of recruiters and hiring managers. A more scientific approach should dramatically improve the quality of hire in terms of reduced time to productivity, lower attrition and better return on investment of recruiting dollars.
The options for predictive analytics in recruiting are potentially limitless, providing that organisations are able to effectively utilise the plethora of recruiting data it already has e.g. data on high, medium and low performing employees; candidate demographics, sources of hire and background data; assessment and psychometric data; structured interview data etc.
Harnessing this data through working with internal or external HR analytics practitioners will enable the evolution of algorithms that predict hiring success, optimal candidate sourcing channels, high potential candidates, flight risk of new hires, cultural fit etc.
I’m not advocating replacing recruiters with algorithms (although some observers believe this will eventually prove to be the case), but the tools are there to make the acquisition of talent more about science and less about guesswork. The combination of great recruiters and relevant predictive analytics could be a potent one for organisations and one that should provide competitive advantage and positive business outcomes.