The strength of weak ties: the network that quietly carries you

In 1973, Mark Granovetter published The Strength of Weak Ties and quietly rearranged how researchers think about social networks. The headline finding: people who got new jobs through a contact rarely got them from a close friend. They got them from someone they hadn't seen in months - an old colleague, a roommate from years ago, a friend-of-a-friend at a wedding. Strong ties point inward at the same information you already have. Weak ties bridge into other clusters where the new stuff lives.

Half a century later the finding has been replicated for opportunities of every shape: jobs, apartments, medical referrals, books read at the right moment. The mechanism is informational - your inner circle reads what you read, knows whom you know, lives where you live. The weak ties don't.

Granovetter's evidence was a survey - suggestive, but correlational. In his sample of professionals around Boston, only 16.7% found their job through a frequent contact; 55.6% through an occasional one; 27.8% through a rare one. Weaker ties clearly did more work, but a survey can't prove the tie caused the job rather than just coinciding with it.

In 2022, a team led by Rajkumar, Saint-Jacques, Bojinov, Brynjolfsson and Aral closed that gap in Science with the largest causal test of the theory ever run: a five-year randomized experiment on roughly 20 million LinkedIn users, two billion new connections, and around 600,000 jobs accepted. By nudging the "People You May Know" suggestions, they could actually measure which kinds of ties caused job moves. The headline held - weaker ties caused more job mobility than strong ones - but with a sharp twist most retellings miss.

The most valuable ties were moderately weak, not the absolute weakest. The job-finding benefit peaked around ten mutual connections and fell off beyond that - a stranger with nothing in common is too far to help, an inseparable friend is too close to know anything new. And it split by sector: weak ties drove far more mobility in digital, remote-friendly industries, while strong ties did better in less-digital ones. For keeping in touch, that pins down the exact band a system has to carry: the moderately-weak middle, which is also precisely the band you forget first.

Why do weak ties carry the new information? The answer is structural, not emotional. Granovetter's term was the bridge - a tie that connects two clusters that would otherwise be disconnected. Sociologist Ronald Burt sharpened this in 2004 into structural holes: the gaps between separate groups. People who span those gaps - brokers - get information earlier, broader, and more diverse than people buried inside a single tight cluster, and Burt found their ideas were rated more valuable and their careers paid and promoted better.

There's a subtle correction in here worth carrying. Granovetter said tie weakness predicts bridging; Burt argued it's the bridging itself that does the work. Practically they agree: it isn't weakness that helps you, it's reach into a different cluster. A weak tie is simply the most common form that reach takes - the old colleague now in another industry, the friend-of-a-friend who moved cities. So when you decide who to keep in touch with for opportunity's sake, favour the people who connect you to a world you're not already in, not just the people you happen to message least. That distinction is the whole point of using a personal CRM for networking.

Granovetter's original split was quantitative: time spent + emotional intensity + intimacy + reciprocal services. The practical translation for a personal CRM is cruder and more honest - strong = you talk to them every few weeks without prompting, weak = you'd be glad to hear from them but you won't initiate, absent = you've forgotten you ever shared a coffee. The system you're missing isn't for the strong ties (you maintain them by gravity) or the absent ones (let them go). It's for the middle band: weak ties that quietly degrade to absent over a few inattentive years.

Memory budgets aren't equal. Brains spend cycles on the people physically near you, the conversations under emotional load, the recurring obligations - kids, partners, parents, the boss. Weak ties are explicitly the people whose absence doesn't hurt today. So they're the first to drop when attention runs short - not because you stopped caring, but because nothing was requesting a memory cycle. The asymmetry between "strong tie that nudges itself" and "weak tie that needs you to remember on your own" is the gap a tool fills.

A personal CRM doesn't conjure intimacy. What it does is push the weak-tie list back into your peripheral vision often enough that one of those people surfaces when you have a free hour. Contact Book does this with a per-person cadence (yearly works for most weak ties), an overdue list on the dashboard, and a one-line conversation log so the next message has a hook to open with. The mental model is groundskeeper, not curator.

There's a failure mode worth naming. People with a personal CRM and a weak-tie list sometimes start sending "checking in!" messages on schedule, with the warmth of a Slack reminder. Receivers can smell it. The point of yearly contact isn't to prove you remembered; it's to actually have a moment where the other person crosses your mind and you say something specific. If you can't write something specific, the cadence is wrong, the relationship is genuinely over, or you're treating the tool as an obligation list. None of those are failures of the person (see also: the etiquette of the follow-up).