Recruitment has changed a lot over the decades. What used to be done via print classified ads and fax machines is now, for many businesses, almost entirely digital—with even job interviews sometimes taking place over Skype. In some ways, recruiters’ jobs have become more difficult as a result of these technological shifts: it’s now imperative that businesses develop and track their employer brands across numerous online channels, meaning that recruitment is no longer a matter of crafting a succinct job ad and sifting through a stack of resumes.
At the same time, technology has the power to make recruiting more effective overall (even in the midst of a serious talent shortage in many parts of the globe) and make life easier for recruiters in a handful of ways, from optimizing audience selection for candidate sourcing to weeding out unsuitable resumes. What technology is making all of this possible? Machine learning.
What is Machine Learning?
Machine learning (ML) has become something of a buzzword in the past several years. But now that the blockchain is the hip, misunderstood technology of choice, people are finally starting to dig into what ML actually is and what its uses are. So what is machine learning? Simply put, it’s the process by which computer algorithms “learn” to draw connections and correlations from large quantities of data. This is how spam filtering in e-mail inboxes works. A machine learning algorithm is presented with significant quantities of e-mails (more than a human could conceivably process), some of which have been flagged as spam by human users, and over time the machine gets better and better at identifying the features of a spam e-mail and filtering them out of your inbox.
What’s crucial here is that programmers don’t need to “teach” the program what spam or promotion e-mails or messages from social networks are—it’s able to figure these things out on its own, and in that way, it develops a granular understanding of how to recognize various types of e-mail based on nuances in things like capitalization and punctuation that humans might not immediately identify as the hallmarks of a particular category of message. It’s this same logic that is increasingly making it possible for computer programs to identify images, perform sentiment analyses on music, and optimize processes across a wide variety of industries. As the sheer amount of data available on the internet continues to grow, machine learning algorithms will be increasingly empowered to uncover hidden correlations that make it possible to perform tasks that, not long ago, would have required human labor.
Smarter Candidate Sourcing
This is all exciting from a theoretical perspective, but what implications does it have for today’s ever-evolving recruitment field? As we alluded to above, there are a lot of tasks that come along with the modern recruiter’s job description—and some of them are prime candidates for the kind of automation that ML makes possible. For instance, when it came time to narrow down a list of candidates from a large number of resumes, machine learning processes could potentially learn to weed out some unsuitable candidates based on similarities in their CVs or applications. Likewise, these advanced algorithms could also gather data from thousands of recruitment marketing campaigns in order to suggest ideal ranges for things like ad budgets and targeting options.
Let’s take a close look at what the above example of machine learning for recruitment could look like. Imagine that you’re trying to source qualified web developers for an open position at your business. You decide to run an Instagram ad campaign targeted at your candidate personas. Though you have a fairly clear idea of the level of experience your ideal hire would need, not to mention their culture and workstyle preferences, the platform gives you a host of targeting options that go far beyond what you had conceptualized for your personas. Instead of guessing or spending an inordinate amount of time on further demographic research, you could let a machine learning algorithm trained on hundreds or thousands of previous ad campaigns recommend a target market based on the minutest details of past successes.
When it came time to specify a budget for the ads, the algorithm could again automate a lot of the guesswork. Sure, you might have a particular ads budget in mind going into a task like sourcing candidates through social media, but machine learning would be able to give you a sense of how effective that ad budget would be, letting you know if you were in danger of over- or under-spending on a particular campaign. In this way, you would be able to make sure that your outreach was effectively serving your goal, rather than potentially running a campaign that was aimed at the wrong audience or not optimized for cost—both of which could potentially slow down hiring, increase the cost per hire, and slow the growth of your talent pool.
Looking to the Future
Though machine learning is still a developing technology, all of what we’ve described above is out there being put to hard use by recruiters. As this technology continues to grow and evolve, new applications will no doubt come to light, leading to even more automation in the recruitment space. Picture a machine learning workflow that identifies potential candidates that it determines are likely to be ready to switch jobs and proactively builds files for them in your applicant tracking system (ATS). The machine learning algorithm then ensures that those users are targeted in any future recruitment marketing campaigns, follows up with them via automated text or e-mail messages if and when they apply, and uses data from the interview stages of the process to help hiring managers to make the right choice.
This level of automation is still a ways off, but to the extent to which ML is already being adopted in recruitment workflows throughout the industry, it’s having a meaningful impact. At the beginning of this post we alluded to the ways that technology has added complexity to the task of building a high-quality team for your company—but the axe swings both ways. Just as technology can and does add complexity, it can also streamline processes like candidate sourcing, helping those in the recruitment sphere to do their jobs better and more efficiently.