Subscribe to DSC Newsletter

Leveraging Predictive Analytics to Avoid a Major Point of Hiring Failure

By Greta Roberts, CEO, Talent Analytics, Corp. @GretaRoberts

Program Chair, Predictive Analytics World for Workforce

What is an employer’s most business-critical corporate process? At or near the top of this list has to be hiring employees that deliver more value to their role and company than they cost to their employer. Employees bring in revenue, rescue a customer, make your products, deliver goods, and sustain your profitability going forward. Identifying the right people, and avoiding the wrong ones, is an imperative to business sustainability.

But how is initial candidate screening handled at your organization? Many employers take an approach that isn’t at all what we expected.

Talent Analytics uses predictive modeling to help organizations reduce attrition and increase on the job performance with in high volume job roles. Our projects give us insight into the entire hiring process.

We’ve discovered that a surprising number of companies, typically midsize and large enterprises who screen a large volume of candidates, often relegate this single most critical task to individuals far removed from the line of business.  At times this task is given to contractors, interns, temps or external part-time employees.

What a huge point of failure leading to massive hiring errors and missed opportunities.

Organizations need to replace this manual, error-prone screening process with predictive analytics processes and technologies that are unbiased, trained to look for a specific combination of predictive factors.  Predictive models don’t get tired; they learn and get better over time; they equally weight candidate.

Predictive Analytics in the Hiring Process Can Take Some Getting Used to

We’ve uncovered two truths when recommending hiring processes including talent analytics. First, some people seem to inherently distrust the analytics-based process. And second, those who distrust it are often the same people who consign candidate screening to under qualified screeners.

We understand hesitation. Including predictive modeling in the hiring cycle is new to many organizations. It can be difficult for many people to get their head around the models and what they’re doing.

But while skepticism of new approaches may be logical, blind reliance on old, ineffective methods is not logical and is not fair. And that’s certainly the case when depending on part-timers or interns to screen job candidates.

These screeners typically need to review a large number of résumés in a short period of time for a wide variety of roles. Even if they’ve had weeks of intense training on how to screen candidates—which is almost never the case—it’s impossible for a single individual to keep all relevant variables in mind as they scan résumés or conduct 10-minute screening phone calls. As a consequence, the initial decision about one of your most important corporate processes is made in what’s clearly an error-prone manner.

Adding Scientific Methods Lead to Fair, Repeatable, Decisions

A far more effective method is to include talent analytics to ensure your hiring processes are based on relevant data. Analytics technology can process high volumes of data, without bias, without excluding important variables, without growing tired—and with consistency, from the first candidate to the last.

The result is a hiring process that’s accurate, that’s repeatable and that’s far more likely to surface the best candidates and eliminate those you want to avoid.

So there doesn’t have to be a battle between traditional screening processes and talent analytics. Just be sure you’re applying the two approaches with the right balance that will deliver the results you want: a workforce that truly supports your business objectives.

Views: 843

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Alex Karman on August 14, 2015 at 3:31pm

Pre-screen enough resumes without human interaction and you will be accused simultaneously of discrimination and reverse-discrimination

Comment by Alexander Kashko on August 22, 2014 at 7:09am

Agreed no one wants to read 2500 resumes. Perhaps the key would be to have people read resumes and flag them as "interview" or not and allow a machine leanring algorithm to devise it s model of what is wanted. Pehraps after say 20 CVs the model would have enough information to filter out clearly unsuitable ones.

The problem as I see it is that carelessly used auto screening can result in all applications being rejected without even being seen by a human. This may be an organisational problem not a technical one.

Another refinement would be to look at the submitted CVs of hired employees after a few years, and see how performance matched the CV and interviewer notes. This could refine the model depending on whether short or long term performance was the criterion

An interesting topic.

Comment by Adam Chiou on August 20, 2014 at 7:14am

A lot of companies are already auto-screening the resumes based on key words.  That practice could introduce bias based on predefined key words.  But no one recruiter wants to read 25000 resumes.  So the analytics model should solve the problems of either hiring the wrong candidate or missing a good qualified candidate by using a particular model.  I suppose a sophisticated resume machine learning model that goes beyond the key word matching could reveal additional insight about a candidate.  Interested to learn more details about Talent Analytics, Corp's talent analytics model. 

Comment by Alexander Kashko on August 20, 2014 at 6:37am

If I were  dealing with hiring I would want to run the analytic method in parallel with human screeners and see  the extent to which the two groups overlapped.  

I would also be a bit worried that the auomated screen would eliminate good candidates who do not fit the standard mould, for example someone with a varied career, or time out for raising children. 

I recall reading of a company that recieved 25000 applications for a run of the mill job and their automated screening process rejected them all. 

Follow Us

Videos

  • Add Videos
  • View All

Resources

© 2017   Data Science Central   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service