--- title: "A Weak Tie Is Not a Weak Relationship" author: "Jon Chu" published: 2026-09-01 canonical: https://www.therevolv.com/signal/weak-ties-are-not-weak-relationships tags: ["Relationship Intelligence", "Networking", "Strategy"] reading_time_minutes: 7 --- # A Weak Tie Is Not a Weak Relationship *LinkedIn tested a fifty-year-old idea on twenty million people. The jobs traveled through acquaintances, but only at a very particular distance.* In 2022, LinkedIn ran an experiment on twenty million people. Most of them never knew they were in it. For five years, the "People You May Know" algorithm quietly varied its suggestions. Some users were nudged toward close connections, others toward distant ones. Two billion new ties and six hundred thousand new jobs later, the researchers were settling a fifty-year-old argument about where opportunity actually comes from, and the results ran in *Science*. [1] The distant connections won. But not the most distant ones. The jobs traveled through acquaintances rather than friends, and the effect peaked at a specific distance: moderately weak ties, around ten mutual connections, moved more careers than either the close friends or the near-strangers. Past that point, more distance stopped helping. That refinement is the part worth keeping, because it argues against instinct instead of for it. If the payoff were simply *the more distant the better*, you could follow a rule. It is not. The value sits at a particular distance, and nothing in your memory tells you which of your ties are standing there. --- ## The Argument LinkedIn Settled You already know the claim. Granovetter, 1973: opportunity travels through acquaintances, because your close ties live in your world and hear the same news you do. [2] By the time something reaches them, it has usually reached you. Strong ties give you trust. Weak ties give you news. But fifty years of retelling flattened the sharp part, and the sharp part is this: > A weak tie is not a weak relationship. It is a measure of distance. Its value has nothing to do with how rarely you talk. It comes from how much of their world you have never seen. And that value is not fixed. It moves whenever the far side moves: a job change, a new city, a market entry, a fundraise. --- ## What It Costs You Here is what the failure looks like from inside a life. You sat next to her at a dinner two years ago. Good conversation, traded numbers, you even wrote a note afterward. Then nothing, because nothing needed to happen. Last month she took a senior role at a company entering your market, and you found out the way a stranger finds out: from an announcement, days late, while a mutual friend one seat over had known for weeks. Nothing about the relationship changed. Everything about its value did. She was never a weak tie. She was a bridge you forgot you owned, and the stale date next to her name was measuring the one axis that does not matter. What you needed was not a reminder to follow up. It was a notice that something on the far side had moved: ```revolv-card { "label": "Revolv · Why this surfaced", "stamp": "6 days ago", "name": "Maya Okonkwo", "role": "VP Partnerships, Arcadia Health · new", "rows": [ { "label": "Was", "value": "Director of BD, Telos Systems (4 yrs)" }, { "label": "You met", "value": "Agora dinner, New York · Nov 2024 · 1 note" }, { "label": "Distance", "value": "9 mutual · no overlap with your current work" } ], "why": "Arcadia announced entry into claims automation on Aug 19. That is the category in your March note on Telos. Maya is your only tie inside it.", "action": "Send a note →", "caption": "A composed illustration of the notice, not a screenshot. What matters is which line does the work: not “you have not spoken in 21 months”, which was true the whole time and meant nothing, but the change on the far side." } ``` The hard part is not finding out that Arcadia entered claims automation. That is public the morning it happens, and a feed will hand it to you along with four hundred other things. The hard part is that the announcement means nothing on its own. It becomes yours only when something connects it to a note you wrote in March, about a company you have not thought about since, belonging to a person whose name you would never have thought to search. Knowing the world moved is cheap. Knowing that this particular movement is yours is the work. Some version of this has happened to you this year. The unsettling part is that you cannot know which tie it was. --- ## You Will Not Fix This With Discipline The instinct is to follow up more. Be better. The research says that will not save you either. A 2011 study in *Organization Science* had executives reconnect with dormant ties, people they had not spoken to in years. [3] Those conversations delivered more novel information than their active relationships did, and the executives had expected less from them than they got. They had been avoiding the one kind of outreach that helped them most, because silence makes a call feel awkward and closeness feels safe. So the deck is stacked twice. Your instincts point at the nearest ten people, and your tools agree with your instincts. Memory sorts by closeness, phones by recency, both toward people whose information you already have. What a network actually needs is two dimensions, and the second one moves. How strong the tie is, and how much world it bridges. Trust lives on the first axis and holds still. Discovery lives on the second and changes without telling you. Every product you can buy, and every instinct you carry, measures the axis that does not move. The interesting question about a network was never "who have I talked to lately?" It is "whose position just changed in a way that changes mine?" Nobody holds that question open across two hundred quiet ties. The executives in the study could not manage a handful. --- ## The Room You Cannot Build Alone Noticing is only half of it. The second axis has two moves, and instinct is bad at both. The first is seeing the bridges you already own. The second is acquiring bridges you do not have yet, and that is where most people quietly give up, because the rooms available to them are the wrong rooms. Your industry's annual conference. The company offsite. Alumni drinks. Every one of them is assembled around the thing you already have in common, which is another way of saying a room with no distance in it. You will leave with names. You will not leave with news, because nobody in that room knows anything you were not going to hear anyway. Now read the dinner again. You did not engineer that meeting. Somebody sat you next to a person whose world did not overlap yours, and two years later that accident turned out to be the only tie you had inside a market you cared about. The accident is the product. Distance does not accumulate on its own, and it does not survive a room that assembled itself. It gets built one table at a time, by someone choosing every seat. That is the problem I am working on with [Revolv](/platform). Not another place to store contacts. An answer to the second axis in both directions: the quiet tie whose far side just moved, and the table where the next quiet tie gets made. Granovetter has been right for fifty years. LinkedIn proved it on twenty million people who never agreed to be proved on. And almost nothing about how anyone keeps a network has changed since, because both moves that follow from it are things a person cannot do alone. So the question is not who you should call this week. It is how many bridges you have already forgotten you own, what changed on the far side of them this month, and who you have still never met because every room you ever walked into agreed with you. --- ## Sources 1. Rajkumar, K., Saint-Jacques, G., Bojinov, I., Brynjolfsson, E., & Aral, S. (2022). A causal test of the strength of weak ties. *Science*, 377(6612), 1304-1310. [doi.org/10.1126/science.abl4476](https://doi.org/10.1126/science.abl4476). The inverted-U finding, moderately weak ties as the most productive channel with diminishing returns to further weakness, is the paper's own refinement of Granovetter. The effect was strongest in industries where work has moved online. 2. Granovetter, M. S. (1973). The strength of weak ties. *American Journal of Sociology*, 78(6), 1360-1380. [doi.org/10.1086/225469](https://doi.org/10.1086/225469). 3. Levin, D. Z., Walter, J., & Murnighan, J. K. (2011). Dormant ties: The value of reconnecting. *Organization Science*, 22(4), 923-939. [doi.org/10.1287/orsc.1100.0576](https://doi.org/10.1287/orsc.1100.0576).