The culture fit worth hiring for is the willingness to fix things.
When I took over a team, I was told during the handoff to keep an eye on the technical quality of the work of one of the tech leads. The warning came without malice, as a request to pay attention, but it was enough for me to form a picture of that person before I'd even met him. I figured he really was a weak tech lead, someone out of date or someone who took far too long to deliver anything.
The picture didn't last long. The whole team had inherited a system nobody knew, with no documentation and no one to ask, and the problem I found was one the whole team faced. In about two weeks, that same person, on his own initiative, doing the work himself, ran sessions with the team, mapped out the technical aspects of the system, shared what he knew, and documented it. The tech lead I'd been warned about ended up as the team's go-to person.
After the first month, I already felt confident leaving technical decisions with him. The attention to his technical work I'd been asked to give never proved necessary, because he saw what the team needed and acted. This remains the clearest example of what I look for when I assess culture fit in a hire.
In every hiring process I've run, behavior and culture came before technical skill. Given a choice between someone with stronger technical skills who struggles with the behavioral side of the job and someone who brings willingness, autonomy, self-management, and a focus on delivery, I'll take the second. Over the years the criterion has proven effective, including with people who needed training and stood out in other ways. I place special value on people who notice something the team needs, act without waiting to be asked, and share what they've gained. That tech lead could have figured out the system and kept the map to himself, like the professional who lives in a cave and only comes out when the task is done.
The handoff warning was about technical skill, and his attitude was what made the difference over those two weeks. In the hiring meeting, that attitude goes by the name culture fit, and the name carries a risk with it, because it has room for that tech lead's willingness and also for affinity, the comfort of approving someone similar to whoever is already in the room.
Amy Kristof-Brown, Ryan Zimmerman, and Erin Johnson pooled 172 studies and 836 effect sizes in a meta-analysis on the consequences of fit at work, and fit scores well on everything that measures how people feel. People who fit the organization come out more satisfied (corrected correlation of 0.44), more committed (0.51), and with less intent to quit (-0.35). When the question is how much people deliver, the association nearly disappears: 0.07 for overall performance and 0.13 for task performance, based on smaller sets of studies. These are correlations, from research published through 2005 and largely measured by self-report, and the association with actually quitting is weaker than the association with intent to quit.
I think of that tech lead when I see that 0.07. What I value in his case is the effect on the team, which came to understand a system nobody had understood.
Lauren Rivera went to look at fit up close, where it gets decided. She conducted 120 interviews with evaluators at elite U.S. investment banks, law firms, and consulting firms, and what they called culture was, in practice, similarity in leisure pursuits, background, and style. That similarity often outweighed productivity, and more than half of the evaluators said fit was the most important criterion in the interview. When they explained why, they talked about turnover: a lot of people leave these firms, and fit was a way to hold on to the people who came in. The explanation is the interviewees' own, and Rivera didn't measure whether it worked.
Taken together, the two studies push me toward a reading that's uncomfortable for my own criterion. If fit goes hand in hand with staying and hardly at all with delivering, and if evaluators, at least in the firms Rivera studied, use culture as a name for similarity, then putting behavior and culture first risks rewarding the people who stay and the people who resemble the team.
My defense is to be careful about what I mean by the word. The fit I look for is observable behavior tied to delivery and to improving the work, the kind of thing that tech lead did with an orphaned system. It shows up in concrete stories about what the person saw missing, what they did about it, and who else had that knowledge afterward, and none of that depends on their sharing the team's tastes or background. Atlassian took a similar path: according to Aubrey Blanche, the company replaced the search for "culture fit" with an assessment of values, in a 2018 account about diversity and inclusion.
Rivera herself shows how far this helps. At the consulting firms, case interviews with structured technical questions reduced the weight of fit without eliminating it, and 40% of consultants still ranked it first. The comparison is across types of firms, with no experiment behind it, and that 40% is a reminder that evaluators need to know what they're looking for before they walk into the room.
Ray Dalio, describing who leaves Bridgewater, writes that when there's a parting of ways, it's usually because the person realizes they can't be excellent and happy there. Read that way, staying requires excellence, and the distinction I draw from it is that some people stay through tolerance, because they can put up with what's wrong. The research sets the bounds of what to expect from fit (satisfaction and the desire to stay), and the advantage of my criterion remains my own bet, with no study demonstrating that advantage.
The hired defect
The fit filter can also reward people who absorb flaws. Imagine someone else facing the same kind of system, with no documentation and no one to ask. They could learn to live with it, figure out where it breaks, and over time become the only one who knows how to restart it. The team would come to depend on them, and the system would remain unfamiliar to everyone else. In an interview, they would tell a good story about adapting, and an evaluator looking for someone who "understands how things work here" and "isn't scared off by what things are really like on the team" would approve them.
The environment's defect would be hired into the role along with them. I call this risk the hired defect: preserving a flaw in the environment by selecting someone for their ability to live with it. I've seen it happen, though it was the exception.
From what I've seen, a flaw survives as long as someone is absorbing it. The broken process keeps running because someone fills in the missing step by hand, and the unstable integration stays up because someone knows how to restart it quietly. The pressure to fix things comes from discomfort, and hiring for tolerance removes the discomfort while leaving the cause where it is. People who tolerate too much may deliver below expectations, since they spend part of their time managing the defect, and that time looks like work in the reports. A team that hires this way grows more and more like itself, patches and all.
Overtime is another way of absorbing the flaw, and in tech it can be mistaken for the willingness I look for. Staying late, on its own, says little about the desire to fix things. When overtime only compensates for the flaw, the cause stays intact, and the team may learn that the problem has a solution as long as someone is willing to stay. I know several people who, when called in to solve a specific problem, fix the flaw at the root in the process, and that takes more work in the moment. The effort can include understanding why the flaw happens, convincing someone it's worth changing, or delaying the immediate response a little.
Fixing problems at the root takes more effort now so the team needs less later, and that's what I bet on when I hire.
Not every tough condition of a role is a defect. Steve Huynh, a former Principal Engineer and Bar Raiser at Amazon, describes in a guide for candidates how companies read role-specific fit signals, such as comfort with ambiguous requirements, and reject people who lack that fit however strong they are technically. Asking for that in certain roles is legitimate. The ambiguity of a product still in discovery is part of the job, whereas the undocumented system that team inherited was a problem someone could fix, and someone did. A person with the willingness I look for asks which of the two cases they're in before accepting the condition, and may find that the condition came with the job.
When someone in the decision meeting says a candidate is a culture fit, my next question is what the candidate has seen missing on a team and what they did about it. The answer I look for resembles that tech lead's story: a system nobody understood and a team that came to understand it without anyone asking. If the answer stops at adapting easily, handling pressure, and not being scared off by a mess, I see the risk of hiring the person who, a few years from now, will be the only one who knows how to restart the system, and who people may speak very highly of at the next handoff.