Most careers turn on an introduction nobody planned

The most important professional relationship of your life probably began with a coincidence you had almost nothing to do with. A dinner you nearly skipped. A delayed train. A friend of a friend who happened to be in the room.

Slightly humbling, when you think about it. It's also something we can influence.

In the largest study ever run on how people find work, the picture flips from luck to design. In 2022 a team from MIT, Stanford, Harvard and LinkedIn ran randomised experiments on LinkedIn's "People You May Know", published in Science, across twenty million people, five years and around six hundred thousand new jobs. The ties that produced the most jobs were the moderately weak ones. Acquaintances. People two steps removed. And the introductions were made by an algorithm.

Mark Granovetter argued exactly this in 1973, in a paper called "The Strength of Weak Ties". Most people who found a job through a contact had found it through someone they saw only occasionally, not a close friend. What was an elegant theory then is now a causal finding at the scale of a small country.

I have spent my career inside that fact without naming it.

The accident you cannot plan

Ask anyone how they met the person who changed their career, and the story almost always starts with some version of "funnily enough". Then they describe the chance encounter, and they tell it like a small miracle.

We call it serendipity and treat it as a kind of magic. Christian Busch, who studies this at LSE and NYU, has a sharper term. He calls it active luck, the kind that shows up far more often for people who arrange the conditions for it. Luck has a method.

A headhunter is someone paid to apply that method on purpose.

Why luck runs out

Left alone, networks get narrower, not wider. McPherson, Smith-Lovin and Cook called it homophily in their 2001 review, one of the most cited findings in all of social science. Similarity breeds connection. We gravitate to people who share our background, our schools, our way of seeing the problem. Pleasant company. Narrow inputs.

The result is comfortable and quietly expensive. Your network fills with people who confirm what you already think. The acquaintance who would have changed your mind, or your trajectory, sits two introductions away and never arrives.

The people who most need a different room are usually the least likely to walk into one by chance.

The loneliest version of this

It gets harder the higher you go, which is not how the promotion was sold to you. RHR International surveyed eighty-three chief executives and found half of them described real isolation in the role. Among first-time CEOs who felt it, around seventy per cent said it was hurting their performance.

The reason is structural. The further up you climb, the fewer people there are who have faced your specific problem, and the ones who have are running their own companies and carrying their own diaries. The pool of people who could genuinely help shrinks to almost nothing, right as the decisions get largest.

I have sat across the table from boards trying to understand why a leader they trusted made a decision with obvious problems in hindsight. The honest answer is usually that nobody in that leader's world was positioned to tell them otherwise. Not because those people were weak. Because the relationship that would have caught it was never built.

What I actually do

Strip away the process and executive search is matchmaking with stakes. I work out who a person needs to meet and why, then engineer a conversation that would almost certainly never have happened on its own.

The good version of this is not transactional. It is understanding someone well enough to know which stranger will matter to them, and being right often enough that they trust the next introduction too.

For years that has been a craft. Slow, manual, bounded by how many people one person can hold in their head.

What changes when a machine can do it

This is the part I find genuinely interesting, and it is why I have started building.

The Science study does more than show that weak ties matter. It shows that a machine deliberately choosing who you meet can move careers at the scale of millions. An algorithm made those introductions, and they worked.

There is a reflex to assume software cannot do the human part. The evidence is starting to complicate that. Researchers at Harvard Business School found last year that people interacting with an AI companion felt less lonely, roughly as much as talking to another person, and more than scrolling. What mattered was feeling heard. People underestimated, beforehand, how much it would help.

Software will not replace the judgement inside a real relationship. It can still lift the ceiling on how many of the right connections one person makes. The deliberate work I do by hand no longer has to be limited to the few hundred people one headhunter can carry.

So I have started building in that direction.

The uncomfortable bit

We treat the best relationships of our lives as happy accidents. It makes a better story. It also lets us off the hook for the ones that never happened.

We have known since 1973 that the connections which change careers are mostly the ones we did not plan. We now have the tools to plan them anyway. Carrying on as though it is all chance is starting to look like a choice.

Most people have never met the person who would have changed their direction. They never will, and they will call it chance.

So here is the question worth sitting with. Who is the person you most needed to meet five years ago, and what would have had to happen for you to actually meet them?

If the honest answer is "pure luck", that is worth thinking about. Luck, it turns out, can be built.

I help boards and founders engineer the connections that matter, in executive search and in what I am building next. Start a conversation

FAQ

Do weak ties really matter more than close friends for finding work?

In the largest causal study to date (Rajkumar et al., Science, 2022, around 20 million users), moderately weak ties produced more job moves than strong ties, with the effect peaking at roughly ten mutual connections.

Can serendipity actually be engineered?

Researchers such as Christian Busch (LSE and NYU) argue serendipity is "active luck", far more frequent for people who deliberately create the conditions for unexpected connection.

Why do senior leaders end up isolated?

RHR International found around half of CEOs experience real isolation in the role; the pool of people who have faced their specific problems shrinks as they rise.

Can AI create meaningful connection, or only transactional matching?

Early evidence (De Freitas et al., Harvard Business School, 2024) suggests AI interaction can reduce loneliness comparably to human interaction when people feel heard, though it complements rather than replaces real relationships.

Sources

• Rajkumar, K., Saint-Jacques, G., Bojinov, I., Brynjolfsson, E., & Aral, S. (2022). A causal test of the strength of weak ties. Science, 377, 1304-1310.
• Granovetter, M. (1973). The Strength of Weak Ties. American Journal of Sociology, 78; and Getting a Job (1974), Harvard University Press.
• McPherson, M., Smith-Lovin, L., & Cook, J.M. (2001). Birds of a Feather: Homophily in Social Networks. Annual Review of Sociology, 27.
• Saporito, T.J. (2012). RHR International CEO Snapshot Survey (n=83).
• De Freitas, J., Uguralp, A.K., Uguralp, Z., & Puntoni, S. (2024). AI Companions Reduce Loneliness. Harvard Business School Working Paper 24-078.
• Busch, C. (2020). The Serendipity Mindset.