Is predictive analytics for hiring simply a buzz word or is there evident of its impact? Now, though, we have emerging evidence that people analytics can successfully be applied to improve hiring. This new category of people analytics is known as predictive hiring – and if you haven’t invested in predictive hiring you’ve already fallen behind.
What is predictive hiring?
Predictive analytics for hiring uses modern data and assessment science to project candidates’ future success based on patterns among current employees. Instead of traditional hiring, which relies on six second resume screens and intuition-based interviews, predictive hiring relies on richer sets of data and smart algorithms to recommend best fit candidates to recruiters and hiring managers. In the same way that your Netflix recommendation feed surfaces movies that you are more likely to like, predictive hiring surfaces candidates who are more likely to be a good fit.
When it’s done well, predictive hiring allows you to focus energy on best-fit candidates, drive business outcomes with data, and engage candidates with authentic experiences.
Several important learnings are beginning to emerge in the transition from traditional hiring to predictive analytics for hiring:
1. Obvious (and often misleading) resume information doesn’t help hire the best people for the job.
Google was famous for collecting all the possible measurable data and test scores they could. When they analyzed actual job performance, they found that “GPAs are worthless as a criteria for hiring… we found that they don’t predict anything,” according to Laszlo Bock, their former SVP of People Operations. Instead, “We care a lot about soft skills – leadership, humility, collaboration, adaptability and loving to learn and re-learn.”
Advances in predictive hiring are increasingly able to measure these more important fundamental skills, rather than relying on the minimally predictive resume data.
2. Cast a wider net to surface best-fit candidates.
Gone are the days of posting on one or two job boards or recruiting at half a dozen “target colleges.” Industry leaders have created objective, data-driven hiring profiles, allowing them to target and surface high-fit candidates that their competition never sees. An extreme example is LinkedIn’s campus recruiting team. Tey Scott, Director of Global Talent Acquisition, said the goal is to never go to a campus again: “We’ve moved from schools to skills.”
As organizations cast a wider net in search of high-fit and diverse talent, they will need ever sharper tools to identify the best candidates.
3. Hiring profiles are unique for every company, even for similar jobs.
Many early analytic initiatives attempted to group similar job titles across organizations. At Koru, we’ve found that the predictors for the same job title vary significantly by company.
How to get started
Predictive hiring can feel out of reach for many organizations. It doesn’t have to be. There are multiple ways to get started while limiting your investment and risk. Lock in some quick wins before expanding your efforts. Start with a very clear and discrete project – usually one or two roles that are high volume and high value to your organization. Pull in someone from your HR analytics team, if you have one, and consider contracting with an outside expert who’s done this before.
- Step 1: Focus on the the business outcome(s) that you’re solving. It’s critical to start with the end in mind and what improvement will make a meaningful difference to your business. Is it performance? Retention? Efficiencies? Diversity? As the saying goes, if you don’t know where you’re going, any path will take you there.
- Step 2: Collect data that could be predictive. Start with your hypotheses and cast a net from there. Don’t fall victim to the trap of “throw all the data in and the algorithms will find magical patterns.” Rarely does that ever happen. Start with the data you already have that you believe carries signal and/or signal-rich data that you can quickly capture.
- Step 3: Run the analysis to test your hypotheses. You can run a regression or correlation in Excel, but this where that internal or external analytics help comes in handy. They may use a basic regression to pick up on strong patterns or signals that exist. More complex machine learning algorithms will be required to boost weak signals and to get smarter over time.
- Step 4: Put data in hands of decision-makers. You must make your findings actionable to capture any value. Otherwise, it’s just a research project. Your decision makers (e.g., recruiters, hiring managers, executives) will also give your findings an important face validity test.
- Step 5: Iterate, iterate, iterate. Predictive hiring is not a “set it and forget it” proposition. It’s a process. You will learn, you will accumulate more data, your algorithms will get more accurate, and your end users will better understand how to drive decisions with your insights.
The only competitive advantage in the 21st century is your people. If you’re already beginning to implement predictive hiring, congratulations, keep going. If you haven’t, the future is now. As the Chinese proverb says, “The best time to plant a tree was 20 years ago. The second best time is now.”
Originally published on TLNT
Josh Jarrett is Chief Product Officer and Co-founder of Koru, the leader in predictive hiring based on what really drives performance. Josh has spent his whole career using data to drive business outcomes at organizations as diverse as McKinsey & Company, the National Park Service, and the Bill & Melinda Gates Foundation.