I know several people who are substantially cleverer than I am. A fair number of them haven’t had the careers that everyone, me included, expected of them at twenty-five. I also know people who were, by their own cheerful admission, merely smart enough, and who went on to build businesses, teams and reputations that have lasted for decades. I spent seventeen years running Vertical Leap watching that pattern play out across hiring rounds and promotions, and it still puzzles me how readily we expect brainpower to turn into results.

I should be careful about what I’m measuring. Wealth and status are a poor scorecard for a life, and plenty of very bright people have spent their intelligence on things that never show up on a balance sheet, so my interest here is narrower: effectiveness at work. On that measure, for most work there’s a level of intelligence beyond which more of it stops helping, and that level is a good deal lower than we tend to assume. Above it, what separates people is judgement, follow-through, persistence and the willingness to do dull work properly, and that’s what we should be hiring for and building in ourselves. In July I wrote that intelligence has always been a commodity, getting cheaper by the year; this piece goes a step further and argues that past the threshold, more of it was never buying much in the first place.

Smart enough for what?

The threshold moves with the task. In mathematics, physics or medicine it sits very high, and nobody should pretend otherwise. For most of management, sales, operations and running a small business, it sits far lower than our hiring instincts suggest. Once someone can understand the problem, solving it depends on other things. Can they decide? Can they explain the decision to the people who have to carry it out, work alongside colleagues who see it differently, and keep going until it’s finished?

Most of us have seen a version of this in our own businesses. One manager produces the brilliant analysis, the one everybody admires in the meeting, while another reaches a sound decision a week earlier, gets the team behind it and has it working by the end of the month. Both understood the problem perfectly well, and only one of them changed anything. Once you can grasp what’s going on, extra intellect buys surprisingly little, and the qualities that turn understanding into results start doing nearly all of the work.

Where everyone is clever, something else decides

The clearest illustration, I think, comes from the places that care most about raw intelligence. Investment banks, trading desks and the big strategy consultancies screen ferociously at the door, with top graduates from top universities, numerical tests and case interviews designed to find the quickest minds in the year group. I’d bet that in every one of these firms the most intellectually gifted member of each intake isn’t the one who makes partner, and that nobody inside would find that surprising.

The reason is almost arithmetic. Once everyone who got through the door is in the top few per cent, intelligence has very little variation left to predict anything with. Promotion ends up running on whatever the firm didn’t screen for: stamina, client instinct, the ability to sell, a feel for the politics, and the habit of doing the grind properly at two in the morning. The firms themselves would say, fairly enough, that they promote on impact, and impact is where those other qualities show up.

Two things sharpen the point. The elite degree also bundles a good deal more than intelligence, from work ethic and conformity to stamina and a brand the firm can put in front of its clients. And what these firms really sell is trust and execution under pressure, which is where the qualities beyond intelligence bite hardest.

Lewis Terman’s long-running study, which followed Californian children with very high measured IQs into adulthood, offers a much older data point. When Terman and Melita Oden compared the 150 most successful men in the sample with the 150 least successful, every one of them had cleared the intelligence bar by definition, and the two groups didn’t differ to any real extent in tested intelligence. What set the successful group apart was persistence in the accomplishment of ends, integration toward goals, self-confidence and freedom from feelings of inferiority.

Where the threshold is high

There’s one piece of evidence that pulls the other way, and it deserves a straight answer. The Study of Mathematically Precocious Youth has followed thousands of people who scored in the top 1% for mathematical reasoning at around the age of twelve. Writing about it after 35 years, David Lubinski and Camilla Benbow reported that even within that top 1%, the top quarter outperformed the bottom quarter on doctorates, patents, income and tenure at leading universities. In that sample, more ability kept buying more achievement.

The threshold idea can absorb this, since research careers in maths and science are where the bar sits highest. What I can’t honestly claim is that the data shows a ceiling there, because it runs out before any ceiling appears.

So here’s where I’ve landed. For most work I’m confident the threshold is low, and the evidence that other qualities matter more above it is strong. For the hardest intellectual work it’s high, and we don’t really know where it sits.

What we reward and what works

Organisations aren’t always good at noticing any of this. Articulate confidence, impressive credentials and quick answers in meetings tend to get more recognition than dependable execution, partly because they’re visible and partly because they feel like intelligence. The person who finishes everything they start, on time and without drama, is easy to take for granted.

If you’re hiring, start with the evidence of follow-through. Look at work the person has actually finished and how good it was, and ask how they kept going when a project turned dull or difficult. Watch how they take feedback, because someone who adjusts quickly will outgrow someone who’s quicker but defensive, and notice whether they recognise uncertainty, since people who know the limits of what they know make fewer expensive mistakes.

The research supports that kind of assessment. When Paul Sackett and colleagues revisited the evidence in 2023, they placed structured interviews, work samples and job-knowledge tests among the strongest predictors of job performance. Cognitive ability kept a role, and they didn’t claim any universal threshold, while conscientiousness (work ethic, attention to detail, follow-through) has some of the best evidence behind it of any personality trait.

The same thinking applies to the team you already have. Before blaming the calibre of your people for a business that’s stuck, it’s worth asking what’s limiting their performance — unclear priorities, a process nobody has fixed, two departments that won’t talk to each other. Those usually explain far more than anyone’s IQ, and the fix is often in how the role is designed rather than who fills it, which is the case I made for designing a role as bounded intelligence.

AI pushes the threshold lower still for many jobs, because it cuts the price of analytical horsepower. When Erik Brynjolfsson, Danielle Li and Lindsey Raymond studied 5,172 customer support agents, AI assistance raised issues resolved per hour by 15% on average, with the biggest gains going to less experienced and lower-skilled workers. Strictly, that’s a finding about task performance, but it raises the question of which abilities grow more valuable once analysis is cheap, and my answer is the work either side of it: framing the right problem and giving it context, then judging what comes back, knowing where our knowledge runs out and acting on it. Anthropic’s study of its own hiring scorecards, published in October, points the same way, with interviewers writing more about autonomy and judgement among technical candidates and less about domain expertise, though it’s one firm’s data and the authors warn against generalising from it. The pull runs both ways, mind, because checking difficult AI output takes real expertise, so the same technology may also amplify the most capable people.

Smart enough is a good place to start

High standards still matter, and so does learning. Intelligence, expertise, judgement and character overlap a great deal, and the best people I’ve worked with had plenty of all of them. We can value our own ability to learn and contribute without treating exceptional intelligence as the entry fee for worthwhile work. Most of us are smart enough for the work in front of us, and so are most of the people we employ.

For employers, that means working out the level a role genuinely needs and then assessing the whole of someone’s contribution above it. So when something isn’t working tomorrow morning, ask what’s actually holding this person, this team or this business back. In my experience the answer is usually something we can fix together, which is a far more hopeful place to start than wishing we were all a little cleverer.