A crystal ball for hiring—Part I: How do you predict the future?
Bad hires are costly
Most articles and studies cite something like 10-15x salary. A lot of that comes from opportunity cost but ask most people who have hired the wrong person, and they'll tell you it's a big problem, regardless of what the "true" costs are.
The average manager only hires the "right" person about half the time
It's basically a coin flip for the average person. There are lots of thoughtful people with smart tools and smarter mousetraps who have great ideas to "solve" the problem. But it's hard. There are skill gaps. No one learns about this in school. Interpreting data is tough. Practice (not to mention good feedback for it) are hard to come by. And for some, there are willgaps. The downside costs of a bad hire aren't felt until later, so you can always prioritize something more urgent. And sometimes, you just need someone to fill the role, so you ignore warning signs during the interview process and just fill the role.
That means hiring has the potential to be a huge drag on value creation
For a typical company…it's a costly problem that has a high failure rate. For a fast-growing company that hires a lot of people, it's even worse. But many people still haven't figured out how to solve the problem sustainably or in a way that works consistently.
Google famously spent 5 years analyzing the hiring problem using their resources and huge data set (the result was 4 interviews, done well, predicted 86% success). If you're reading this, you're (probably) not Google.
What should you do?
Tip #1: Own the problem and decide you want to solve it
Recently, I talked to a fast-growing tech company. The founder had grown the company but made mistake after mistake in senior hires. It was clear to him that his hires didn't work out. But in the same breath, he also said people stuff was a bunch of BS. Sigh. So, decide that the problem is big enough you care, but be sensible on how you do it.
Tip #2: Set a goal for how good you want to be and pick a role you're going to start with
I'm a fan of framing a problem practically and then being data-driven on how you want to solve it.
Frequently people assume that you have to jump in with both feet and do it all. Or they read something (perhaps from a company that is selling their approach) and absolutely HAVE TO do it that way, and if you don’t, you'll be a failure. I don't subscribe to black and white thinking. It's not binary. Your goal is to get better. But you need to get better only at the level YOU care about and is valuable for you.
There's a way to visualize it. Every bit of data you collect on a candidate gives you more confidence that you're making the right decision on hiring. If you start at 50% accuracy, you don't need to get to 100%. Just be better than where you were before. And because the costs of mis-hires are much bigger for more valuable roles, you can decide to be more confident for those, but less confident for others.
To be clear, there probably IS an ideal answer to maximize your confidence. If I absolutely HAD to get it right…I would chose a structured behavioral interviews, supplemented with targeted evaluations of a few personality factors like conscientiousness and intelligence, some really good references from thoughtful people, all done against a clear description of the role to be performed. Oh, and all conducted by experts in human behavior and the business's industry.
But that's hard to do in the real world for every hire. So…
Tip #3: Make a choice on getting better…What, how, or who?
There's a LOT of data and guides and manuals on hiring. Some are really good. Some are meh. There's also a lot of research on what personality characteristics imply about a candidate's odds of success. Some of the studies are pretty compelling. Others look good until you dig in. Some people will give you advice on who should conduct these magical tests . But it’s not always practical. All of the advice boils down to the same thing. Bend the curve upwards. Get more confident and make better decisions by evaluating the right stuff (what), in a better way (how), with more competent assessors (who).
Tip #4: Match the value of the role to the need for accuracy and therefore the effort required to collect the data
As you look at the graph above, it’s clear that collecting more data means more confidence in getting it right. But not all data is the same. Some is very easy and cheap to do. Doing a google search and combing through social media posts doesn’t take much effort and is essentially free. But it also might move the needle on accuracy by 1% (and it might even cause you to make a worse decision). Hiring a search firm to source and screen candidates will cost you a third of the candidate’s salary, not to mention the time spent to manage the search firm. And while you may be better than 50/50 (at least for a good search firm), you probably can’t invest 1/3 of EVERY new hire’s salary. So evaluate the cost/benefit for each WHAT/HOW/WHO decision you make and balance that against the criticality of the role.
If you’re starting with 50% hiring accuracy, I would suggest going for 90% accuracy (probably what I described above, all of which will cost time and money to do) for the most critical roles in the business. Maybe 75-80% for the next set of roles. And then 65-75% (with an aspiration to get higher the more effective and efficient you can make your processes) for all remaining roles.