“Daddy, did you check the weather forecast?” This is the question my daughter asks me almost every morning. I’ll be honest, I don’t usually check the forecast. I choose instead to rely on my own judgment. I look at the sky, think about what the weather has been like for the last few days, or stick my head out the door to see how the air feels.
Unfortunately, this results in my family being fairly regularly caught out by the weather and coming home either frozen to the bone, drenched in rain, or hot and bothered. Now, I hold my hands up and take the blame for this. If I were only less stubborn and chose to check the forecast more often, most of the issues could be avoided. But for some reason, I still don’t check it and the issues continue.
I’m mentioning this shortcoming because I’ve noticed something similar happening across the world of HR. People throughout the HR world, myself included, seem to be falling into the trap of making important workforce decisions based entirely on intuition and personal experience. These reflexive and ‘feelingbased’ decisions are often made even now that powerful workforce analytics and AI tools are available. Of course, these tendencies are completely understandable. In our careers, it’s usually our ability to make good decisions intuitively that got us to the positions we’re in, so it’s only natural that we are conditioned to continue in the same way.
However, while intuitive decision making is always going to be crucial, the precision and logic applied by machine learning and AI is something we can all learn from. As these technologies develop, data gathering and analysis methods are becoming ever-more powerful and can help to inform more targeted, effective change across all areas of business.
Just like the weather forecast, ignoring good HR data and analytics can leave us out in the cold.
But to be able to improve the way we look at decision making, I think it’s important to first understand why we so often overlook workforce data and analytics. I see three main causes.
1. The struggle to keep up with technology
One thing that prevents us from fully utilizing workforce data is that it can be difficult to properly interpret without specialized training in data analysis.
The reason for this skillset disconnect is simple. HR professionals were traditionally trained in interpersonal skills primarily, not in the specifics of data science as this kind of training was not relevant until very recently. Of course, interpersonal skills will always be vital but the business world is constantly moving forward, so we should too.
With such rapid advancements taking place all the time, some lag between technology and training is unavoidable. But what if organizations started to prioritize a more holistic approach to training that covered emerging skillsets including modern data science? Molding our approach in ways like this could be integral to ensuring that HR adapts and stays relevant in a changing business world.
It’s only human to put faith in one’s own ability to make judgments, even if that means sometimes overestimating it.
Looking forward, perhaps it will be best to accept that AI and machine learning have huge potential as assistive tools. These technologies are now capable of augmenting our own analytical skills, bypassing natural flaws and biases in our thinking, and providing powerful, unique workforce insights.
2. Not trusting the data
The second issue I see is that many of us in HR seem wary of placing trust in workforce analytics that actually have the potential to be very insightful. I think that when a concept is new and unknown it’s always difficult to put faith in it, and I’ve been guilty of this personally. “How can a computer truly understand a workforce?” – This is the kind of question I have asked myself often, as I’m sure many others have. For me, the only way around this is to familiarize myself with the new concept, and the best way to do that is by looking at the research.
An illustrative example of eye-opening research is Michal Kosinski’s excellent study on AI and personality assessments. The project found that an AI program only needed to analyze 10 Facebook likes to assess somebody’s personality more accurately than their colleagues. It needed 70 likes to know someone better than their friends, 150 for their family members, and just 300 likes to beat, rather incredibly, a spouse. That’s right, a simple AI program can assess somebody’s personality better than the person closest to them in the world based on a few Facebook likes. Imagine what specialized machine learning can achieve for a company’s workforce, especially when combined with a real person’s intuition and decision-making ability.
3. Misinterpreted ethical concerns
The final major issue I see is ethical concerns, or rather the misinterpretation of them. This certainly isn’t to say that there aren’t any ethical considerations surrounding the use of workforce data. On the contrary, misuse of personal data is a genuine problem, and things like the Cambridge Analytica fiasco must be avoided in the future. Risks such as reinforcing particular biases of gender, ethnicity, age, and so on are very real, so should be in the front of any data scientist’s mind at all times.
Fortunately, the world does seem capable of adapting to these potential threats. This is perfectly demonstrated by the recent GDPR laws in Europe, as well as the general increase in public discourse surrounding data protection and misuse. I believe that moving forward we should avoid exclusively talking about the possible negatives without also considering the benefits that workforce data can present.
With real, accurate insights into their people, businesses across all industries are able to make better-informed, more effective decisions. Simply ignoring the available data runs the risk of ending up with subjective, biased, and harmful decisions being made, which can negatively impact any business. Because of this, I feel it is crucial for us to use all available data and technology to help inform the balance between business success and employee wellbeing.
What if, instead of rejecting the power of people analytics, HR departments open the discussion on how to harness workforce data and guide it to be used for good?