How An Algorithm Can Replace a Recruiter (And How It Can’t)
Can an algorithm replace a recruiter?
This question was recently posed by Andrea Woolley, a project director at SAP, and Matthew Jeffery, VP – Head of Global Sourcing and Employment Branding for SAP, when they revealed SAP’s new approach to university recruiting for sales.
What SAP did to recruit new grads for sales
SAP replaced the traditional recruitment techniques of reaching out to a few top universities and schmoozing administrators and professors with what they called “the democratization of university recruiting.”
They opened up their university recruitment to all schools by using an assessment process composed of:
- a 10-minute online cultural fit test
- a 20-minute online situational judgment test
- a day-long in-person assessment bootcamp
What SAP found by using pre-hire sales assessments
- Number of people starting the assessment: 50,000+
- Dropouts in application: from 93% (old way) to 25% (new system)
- Candidates that said the online tools boosted their motivation to continue the application: 75%
- Candidates that said that the bootcamp provided a great insight into working at SAP: 100%
- Complaints about candidate experience: Zero
- Projected cost saving: $372,136 (year one)
- Number of new hires: 500 globally
Safe to say, their new sales recruitment process was a runaway success.
What SAP concluded about their new sales recruitment process
So can an algorithm replace a recruiter? Yes. We have shown that. Technology blessed us with a high potential graduate pipeline that has demonstrated a quality-of-hire prowess through an ability to hit quota, faster than previous graduates, and at higher levels.
How an algorithm can replace a recruiter
After SAP opened up their university recruitment, they received more than 50,000 applications for sales positions. Even if a recruiter spent only 6 seconds per resume, that’s still more than 83 hours spent just on resume pre-screening. Yet SAP saved more than $370,000 in costs a year with their new recruitment system. How?
By replacing manual pre-screening of resumes with automated online assessments and saving their recruiters hundreds of wasted hours.
Before you think eliminating manual pre-screening of resumes is a bad thing, consider that Google – a company that receives 50,000+ resumes a week – admits that a person’s resume doesn’t predict his or her future job performance.
But surely a human being is better at assessing another human being – with all our complexities and quirks – over a cold and impersonal statistical algorithm? Nope.
Research featured in the Harvard Business Review found that compared to using expert human judgment, an algorithm increases the accuracy of selecting successful job candidates by more than 50%.
In fact, what the research has found over and over again is that when experts add their judgment to a data-based algorithm, the outcome generally becomes worse than just using the algorithm alone.
How an algorithm can’t replace a recruiter
Okay, algorithms vastly outperform humans in accurately predicting which job candidates are more likely to succeed. So why not automate the entire hiring process?
I have a confession to make: that’s exactly what I thought we could do at Ideal Candidate. Armed with the data, I was confident that we could create a system that eliminated the biases and inefficiencies that plague sales hiring. And I think we’ve done a good job so far.
But what I didn’t anticipate was that job candidates like (and sometimes need) to have a human touch. That’s why we created the Talent Scout position: we needed someone to engage with our job candidates to help set up interviews and keep them in the loop, and to get feedback from employers about job candidates in order to improve our job matching system.
Because we give job candidates so many opportunities to give us feedback and ask questions, the Talent Scout, the Product Manager, and I have all talked to candidates over email and on the phone. That’s a lot of human touches for an “automated” recruitment process. That’s when we realized that Ideal Candidate needed to be more like a recruitment-as-a-service or a software-with-a-service than a traditional SaaS product.
And candidate engagement is where a skilled recruiter’s true value comes in. Not wasting their time pre-screening hundreds (thousands!) of resumes.
As MIT Sloan School of Management’s Andrew McAfee predicts:
…I’m very confident that data-dominated firms are going to take market share, customers, and profits away from those who are still relying too heavily on their human experts.
The trick is knowing when you can replace human experts with an algorithm and when you can’t.
The data make it clear: an algorithm can replace a recruiter when it comes to assessing job candidates but it can’t when it comes to engaging them.
Are you using an algorithm or a human to assess your job candidates? Let me know in the comments or tweet @recruit_smarter.
Latest posts by Ji-A Min (see all)
- The Recruiting Software Stack: A Buyer’s Guide To The Top 4 Innovations - December 2, 2016
- AI Recruitment: The Future of Automated Recruiting - November 22, 2016
- Google Cloud Jobs API: 4 Things You Need to Know - November 17, 2016