Glassdoor just compiled a list of top companies that are dropping college degree requirements.
There are a couple of different ways of looking at this.
One is to say that the job market is so hot—in favor of employees—that companies are relaxing standards just to get people in. Another way to see it is that the correlation between having a college degree and being a great employee is no longer very strong.
Or, perhaps more accurately, the correlation between not having a degree and not being amazing, is not strong. In other words, there is top-tier talent out there that got there without going to university.
When you look at people who don’t go to school and make their way in the world, these are exceptional human beings. And we should do everything we can to find those people.
Laszlo Bock, Senior VP of People Operations at Google
I think both are likely true, but I side towards the latter, and I think there’s more evidence for it as well.
Google did a massive internal study a number of years ago that looked at what predicted success in the company. They wanted to know if it was whether someone went to college, whether it was the eliteness of the college, their GPA, etc. They found very little correlation between any of it and performance. And that included how well people did on their brain-teaser questions, which they officially abandoned shortly after.
Our own internal research of over 400 graduates found that screening students based on academic performance alone was too blunt an approach to recruitment. It found no evidence to conclude that previous success in higher education correlated with future success in subsequent professional qualifications undertaken.
Maggie Stilwell, Managing Partner for Talent at EY
To me this signals a larger change, which is away from traditional markers of predictive quality and towards a data science approach to hiring. This is what machine learning is for. There are too many variables for humans to capture them all in hiring, and interviewers are famously plagued by biases that cause them to hire the wrong people and ignore the best ones.
It seems to me inevitable that—for better or worse—we’re moving towards using machine learning for this entire process.
We don’t know what makes someone successful or not. We’re horrible at knowing this. And we are not just bad at it—we also think we’re good at it, which is a nasty combination.
Once we have the data required, which I’m sure some people like Google already do, we’ll be able to look at:
- Who has been retroactively marked as successful by the business?
- Who has been retroactively marked as unsuccessful by the business?
- What characteristics do these groups share?
- Does this current candidate that I’m looking at have these characteristics?
And hiring will be based on this going forward. The whole game will be gathering as much data about the candidate as possible to determine how likely they are to succeed.
If this sounds frightening to you, you’re really not going to like the world of big data + machine learning.
I think a big part of our future tech world will be having an up-to-date data bundle on ourselves, full of details, attributes, context, etc., and all of that data will be labeled and tagged according to data type and sensitivity.
So when you ask to be seen by a doctor, or when you apply for a job, or sign up for insurance, you’ll be prompted for the latest data bundle on yourself.
You’ll grant it, but what you give them will be restricted to what they need for their particular use case.
Anyway, this is where hiring is going.
Companies don’t care about credentials anymore. They’re a primitive way of doing what machine learning can now do far better.
Companies want people who will succeed in a particular role, and that’s what data science will tell them.