Networking Without the Cringe

“Networking is fake.” I hear a version of that from people who are otherwise good at their jobs — people who’ll negotiate a vendor contract without blinking and then go quiet at the phrase “grab coffee and pick your brain.”

They’re not wrong about the version of networking they’ve actually run into.

The LinkedIn message that opens with a compliment and pivots to an ask by the second sentence. The “let’s connect!” after a conference panel that never turns into an actual conversation. The group chat that only exists to trade favors. All of that earns the reaction it gets, and I’m not going to spend a paragraph defending it.

What I want to defend is something narrower: asking someone you don’t know well a specific question, once, and meaning it. That’s not a performance skill. It’s closer to a research habit, and it happens to be backed by one of the better-replicated findings in the sociology of how people actually get hired.

The people who actually help you are the ones you barely talk to

In 1973, a sociology graduate student named Mark Granovetter surveyed professional and technical workers in Newton, Massachusetts, about how they’d found their current job through a personal contact. He asked how often they saw that contact: often, occasionally, or rarely. Fifty-six percent had gotten the lead from someone they saw only occasionally. Another 28 percent came from someone they saw rarely. The rest — sixteen percent — came from a close, frequent contact.

Granovetter’s explanation is the part that stuck, not just the numbers. The people you see often — your close friends, your current team — have mostly been marinating in the same information you have. A contact you see rarely has been marinating in someone else’s. He called the finding the strength of weak ties, a deliberately backwards name for a real effect: the tie is weak, which is exactly why it works.

Your closest contacts are the least likely to know something you don’t.

Fifty years later, someone ran the actual experiment

Granovetter’s study was a survey, which can show a pattern but can’t prove what caused it — maybe people who are already good at their jobs are also better at collecting weak ties, and the ties themselves aren’t doing any work. In 2022, a team led by Karthik Rajkumar and Sinan Aral got around that problem the way only a company that runs the network can: they randomized it. LinkedIn’s own “People You May Know” algorithm decides who gets recommended to whom, so for five years the researchers nudged it to vary, at random, how many weak versus strong ties different users were shown — more than twenty million people, over two billion new connections, six hundred thousand new jobs. I’ve read that scale four times and it still doesn’t quite land for me; it reads more like a national census than an experiment, and there’s something I haven’t fully made peace with in a hiring outcome for that many people having been quietly steered by an algorithm none of them knew was running. I don’t have a tidy answer for that, so I’ll leave it here instead of resolving it. What the experiment did confirm, closely, was Granovetter. This time the evidence was causal, from a randomized test rather than a survey people filled out about their own choices. It came with one correction: the very weakest ties, strangers with almost nothing in common, didn’t help much either. The sweet spot sat in the middle, at connections with roughly ten mutual contacts between you, according to the paper.

The effect showed up strongest in software and other digital-heavy industries — the same sectors where sending someone a LinkedIn message is already a normal way to start a conversation.

The list of people worth messaging already exists

You don’t need to meet new people to use any of this. You need to stop ignoring the ones you already know slightly.

Open LinkedIn and scroll past the names you’d message without thinking twice — you already know those people too well for this to work. Slow down at the ones that make you pause: someone from a team you rotated through for one quarter three years ago, a person you sat next to at a single conference, a former manager’s former manager, a college acquaintance who ended up somewhere you’re curious about now. Ten names like that is a real list.

This isn’t about being extroverted. It’s about being specific.

The event with a talk beats the open bar

A LinkedIn connection is a data point. A person you’ve sat in a room with for forty-five minutes is most of the way to a relationship, and it gets there faster than any number of messages will.

Luma and Meetup have absorbed what used to be scattered across a dozen separate mailing lists, which makes the filtering easy: look for the event with a talk on the agenda — a cash bar and a room full of strangers is a much harder room to break into.

A talk gives you forty minutes where you don’t have to make conversation, and it hands you an opening line for free afterward — you ask the person next to you what they thought of the point the speaker made, and it’s a real question, because you actually want to know.

Small events beat big ones for the same reason a weak tie beats a stranger. Fifteen people who all showed up for the same narrow reason already have more in common with you than three hundred people who came for the food.

A first message doesn’t need an ask in it

This is roughly what one looks like, sent to someone from that list of ten:

Notice what isn’t in it. There’s a specific, checkable shared reference — the platform migration, not “I remember working with you” — and an exit that costs the reader nothing to take. If there’s an ask, it comes in the second message, after they’ve replied. By then it isn’t cold anymore.

The warm intro, and the follow-up nobody sends

A warm intro is the same move, routed through someone else. Skip asking your contact to vouch for you — ask them to forward something you already wrote instead, so the only work left for them is one click.

That’s a smaller favor than it looks like, and small favors are the ones that actually get done.

The part almost everyone underperforms on is the follow-up. You meet six people at one event, mean to write to four of them, and by Friday you can’t remember which four or what you meant to say. If one of those threads turns into an actual reply — a recruiter, an intro, a real lead — it usually lands mixed into the same inbox as everything automated you already get, which is the specific mess Vigil exists to sort through. The outbound side, the messages you send out, is still on you to track. The fix for that part is embarrassingly low-tech: write the follow-up note on your phone before you leave the building, while the specific thing they said is still specific enough to use.

The best rule I know for any of this didn’t come from a study. Someone told it to me early, and I’ve never improved on it: any time a person tells you something genuinely useful, say so, specifically, right then. Naming the actual thing that helped lands differently than a generic thanks on the way out the door. People remember being told they mattered to someone’s afternoon. It’s the only part of this that doesn’t need a script.