Evidence
The strength of weak ties
Two sources, and it is worth keeping them apart. Mark Granovetter set out the theory in The Strength of Weak Ties in the American Journal of Sociology in 1973. The interview data sits in his book Getting a Job, published in 1974: of the men in Newton, Massachusetts who found work through a personal contact, the majority had heard about it from someone they saw occasionally or rarely rather than from a close friend.
The mechanism is information overlap. Your close friends know the same openings you do, because they move in your world. An acquaintance sits in a different world and knows things you cannot know. The tie is weak, and that is precisely what makes it useful.
Evidence
Tested at scale in 2022
This stayed a well liked theory with modest data until a large experiment on LinkedIn, reported by Rajkumar and colleagues in Science in 2022, covering roughly 20 million users over five years of changes to how connections were recommended.
Moderately weak ties produced the most job mobility, and the very weakest and the strongest produced less. That refines Granovetter rather than confirming him wholesale: the useful contact is not a stranger, it is someone you genuinely know slightly, with a handful of mutual connections. It also means the productive move is not collecting contacts. It is reactivating people you already know a little.
Claim
What to actually send
The message that works is short, specific, and asks for information rather than a favor. Nobody can grant you a job. Plenty of people can tell you what their team is hiring for and what the work is really like.
So: remind them where you met in one line, say in one line what you are working on now, then ask one answerable question. Whether they are hiring, who does the thing you are curious about, what the market looks like in their part of the industry. No attachments, no request for a call as the first move, no paragraph about your journey.
Action
Send three messages today
Pick three people you have not spoken to in over a year and genuinely know: a former colleague, someone from a course or a conference, a supplier or client you liked.
Write four sentences each, following the shape above. Send them. Then put a note in your calendar for six weeks out to do it again with three different people. Three messages is the whole task, it takes eleven minutes, and it is the single highest yield hour in this course by a wide margin. Expect one reply out of three and do not read the silence as rejection: people are busy, and the two who ignore you this time will sometimes answer in a year.
Example
Craig, a warehouse operations manager, sent four of these on a Sunday evening. Three replied. Two were friendly and useless.
The third was a woman he had worked with six years earlier who had moved into logistics consultancy. She was not hiring. She told him which three companies in the region were rebuilding their distribution networks and that one of them had lost two managers in a month. He applied to that one on Monday through the ordinary careers page, mentioned her name, and got an interview in nine days. She did not refer him. The useful thing she gave him was information he could not have found.
Caution
The way this goes wrong
Two failure modes. The first is asking for too much in the first message: send a CV and a request for a referral to someone you have not spoken to since 2019 and you have made the interaction expensive, so they leave it in the inbox.
The second is contacting people only when you need something, which they notice by the third time. The repair is cheap and slow: send something useful occasionally when you do not want anything. A relevant article, an introduction between two other people, one line congratulating someone on a new role. Ten minutes a month, and the network is still warm when you need it.
Aside
The uncomfortable part of this evidence
If jobs travel along social ties, then hiring favors people with large and varied networks, and network size tracks class, schooling, geography and confidence. That is one of the mechanisms by which advantage reproduces itself, and it is a genuine argument for structured public advertising of roles rather than a system that rewards who you drank with.
I am telling you to use it anyway, because the individual and the policy question are different. If you hire, though, notice the same finding cuts the other way: a pipeline built on referrals will reliably reproduce the team you already have.