Machine Learning Trends for HR

Machine Learning and HR

There have been a few technological advances over the last decade or so that have been pitted against each other in a hypothetical battle for the future of business. One that holds the most potential for real-world applications in HR is machine learning.

Machine learning describes a concept where a machine (basically a piece of software, albeit taking advantage of serious leaps in computational powers of modern hardware) “learns” over time, using previously made calculations and decisions to speed up its future calculations and decisions and make them more successful.

Unfortunately, as is often the case with new technologies, the whole story gets muddled with unrealistic expectations and promises of things that will either never come to life or that will require decades to show any applicable potential.

That is why it is essential to identify machine learning trends that have the potential to be applied to HR as we know it in the foreseeable future and those that do not. Today, we are interested in the former.

Machine Learning and hiring

Let’s say a major corporation receives ten thousand of resumes every year. Let’s say that they make a thousand hires every year. Let’s say that 500 of those work out and 500 do not.

Let’s say that this large corporation keeps all the data surrounding these 10,000 applications and 1,000 hires.

They keep track of who saw the job ad where. They keep track of all the applicants’ resumes and they come up with a way to categorise all the data contained in those resumes. They even include applicants’ social media activities in the data they obtain and keep. They keep track of their standardised, structured interview process. They keep track of the language used in the correspondence. They keep absolutely every piece of data that pertains to the applicants.

They feed all this data into software that uses machine learning, and they feed it continuously from the first day. Soon enough, certain patterns emerge.

The software discovers that a certain job ad website yields more successful hires. A certain interviewer is better at identifying the right talent than other interviewers. People who use a certain type of social media turn out to be better employees. The possibilities are endless, especially when you factor in combinations of individual factors and patterns.

A software that utilises machine learning is the only kind of “entity” that can hope to analyse all this data and find the patterns. A human HR professional could never do something like this. A “traditional”, coded piece of HR software could never do this.

There are already companies that do this and provide this kind of analysis to their clients.

It should be pointed out, however, that some of the patterns and tendencies will turn out to be false positives and that a human will need to have the final decision when all is said and done. This kind of advanced analysis and pattern recognition can greatly increase the success rates of hiring decisions.

Finally, we should also mention the fact that many HR experts are agreeing that machine learning and software based on this principle will be the best way to take out human bias from hiring decisions, improving the quality of the decisions.

Employee attrition

Employee attrition and the subsequent employee turnover is a topic that is being hotly debated these last couple of years, as true costs of high turnover have become widely-known and as companies are trying to do everything in their power to stop this bleeding of talent.

The problem is that even when company’s internal communications are done the right way, it is impossible to do any kind of comprehensive analysis of people’s statements, questions, intentions and decisions that would lead to employee attrition. At least it is impossible for a human HR professional.

However, for a piece of software based on machine learning, certain patterns become identifiable. Just as an example, certain responses on employee satisfaction surveys and drops in efficiency can be observed as precursors to employee attrition and their quitting. There are myriad such signals and they often become important in certain combinations that are impossible to figure out for a mere human being. Machine learning software is already able to do this (to take it a step further even) and the chances are it will only become better at it.

The mythical engagement

The world of HR has been in uproar about employee engagement for quite a while, with people clamouring that these are the Last Days of Engaged Employees and that there is nothing companies can do to prevent their employees from leaving, or at least, mailing it in.

Employee engagement will always be a human-to-human practice; there can be no doubt about that. However, there is plenty to be gained from smart use of machine learning and software that helps identify trends when it comes to engaging employees and preventing them from leaving for “greener pastures”.

Thanks to the increasing use of machine learning in natural language processing, there are companies, like Glint and Workometry  software solutions provide employers with continual insights into how their employees are feeling about their workplace and how engaged they are feeling. This goes a few steps further than simply comparing employee satisfaction surveys. Thanks to new machine learning-based software, it is possible to get insight from seemingly unimportant statements that were previously considered to be too vague or emotion-based to be analysable in any realistic way.

Closing word

Machine learning has made some enormous strides over the last couple of years thanks to certain technological advances, but it is safe to say that we have yet to see its full impact on the world of business and HR specifically.

The important thing is not to go luddite immediately and see it as a harbinger of doom. The future of HR will most probably involve a human-machine collaboration and that can end up being a good thing.

Also read the other articles of Nate Vickery on our site:

No more articles