How to reconnect with an old friend once the first message is sent

If the message itself is what you have been putting off, that has its own article. The short version: in experiments by Liu and colleagues, people who reached out after a gap underestimated how much the recipient appreciated it, and the error was widest when the contact was most surprising, which usually means after the longest silence. The experiments measured single moments of contact, not friendships over time. They say the message will land better than you fear, not that the friendship will come back.

That second part is this article’s question, and the decay research shows why it is a different one. Roberts and Dunbar followed people’s personal networks for eighteen months and found that emotional closeness to friends fell steadily as the time since last contact grew, while closeness to family held across the same gaps. It is observational, and people may simply contact the friends they already feel close to, but the direction is plain: a friendship is largely made of the contact itself.

So the message does not restore the friendship. It restarts the contact, and the friendship comes back only if the contact continues. Most attempts to reconnect fail here: two people exchange a warm summary of the missing years, agree they must do this properly some time, and go back to silence feeling slightly better about each other. Everything below is about getting from the first message to a second one with a reason behind it.

People also search “old friends who have become strangers”, and it is more accurate than it sounds. In a Dutch panel study, Mollenhorst and colleagues interviewed 1,007 adults about their personal networks and re-interviewed 604 of them seven years later. Only 48% of the original network was still in it, and for the closest roles the figure was around 30%. It is one panel in one country, with attrition between the waves, but it is the clearest measurement of how far an ordinary life moves in a few years.

Your old friend has lived through the same turnover. The people they talk to, the job, perhaps the city and the partner have changed since you last knew them, and the version you remember is out of date. That is not a reason to hold back. It is a reason to treat the first conversation as getting to know someone you have a head start with, rather than as a test of whether the old friendship is still intact.

It also means the conversation has to go beyond the summary, and the evidence says that is less awkward than you expect. Kardas, Kumar and Epley assigned people to shallow or deep conversations and asked beforehand how awkward each would feel. The deeper ones felt less awkward and produced more connection than predicted, and the misprediction was largest for the deepest topics. Those were mostly strangers in a lab, so this is about the direction of your forecast, not a promise. Boothby and colleagues found the companion error: people underestimated how much their conversation partners had liked them, even among dorm-mates followed for months. The replay afterwards, where you decide it was awkward and they were being polite, is that bias at work.

What makes a reconnected friendship feel like a friendship rather than a reunion is the sense of being known. Reis, Clark and Holmes define intimacy as feeling understood, validated and cared for, and the word doing the work is perceived: what counts is what the other person registers. The framework rests on a largely correlational literature, so it does not prove that remembering causes closeness. It does explain why “how did the move to Leeds go in the end?” does something “so, what have you been up to?” cannot. The first shows you kept a piece of their life; the second asks them to summarise it.

There is a stranger line of evidence for the same point. Davidson and colleagues studied adults with adult-onset amnesia and found they had made very few new close friends since it began; being unable to return to a previous conversation appeared to block the bonding process. It is a small clinical sample, and amnesia disrupts work, independence and routine as well as memory, so this is an association in an unusual population, not a proven mechanism in a typical one. But it points at the same link: the second conversation is where retained detail either shows up or does not.

In practice this is two moves. Pick up a thread from before the gap: the thing they were worried about, the plan they had, the person they were seeing. Then notice one thing that is coming up for them and write it down afterwards with its date. Follow-up questions help as well. Huang and colleagues found that asking them made people better liked, with perceived responsiveness as the mediator, though one co-author faces data-integrity allegations and the effect has been contested in print, so treat it as corroboration of the responsiveness framework and nothing more.

“Is it worth reconnecting with old friends” is its own search, and the research gives a partial answer. Hall asked 355 adults who had recently moved city, plus a student sample, to estimate the hours spent with new acquaintances and how close they had become. Acquaintance to casual friend took around 50 hours, friend about 90, close friend 200 or more. The thresholds rest on remembered hours from a single lab, so treat them as illustrative. The direction is what matters: closeness is built from accumulated time, and there is no shortcut.

Nobody has measured what reviving a lapsed friendship costs, so calling it cheaper would be inventing a number. What the studies suggest together is narrower. Shared history is not enough on its own, because closeness to friends fades without contact. But a new close friend costs something like two hundred hours of leisure time, and the old friend is the one person for whom those hours have already happened. The question is whether you both want to resume, not whether you can afford to start from nothing.

There is also a case for not reconnecting, and the evidence does not override it. Some friendships ended for reasons that have nothing to do with drift, and the studies on surprising contact were about silence, not a falling-out. If the gap has a cause, the research here does not say how to repair it. If the gap is only a gap, which is how most ties end, it says the door is more open than it feels.

Send one short message with a specific memory or a specific question in it, and no apology for the silence. If they reply, suggest one concrete thing with a date on it: a call on Thursday evening, a coffee the week after next. The friendship was lost to lost contact, so the repair is a next contact, not a long message.

In the first conversation, ask about the threads you remember and let it go past the summary of the years; it will feel less awkward than you expect. Afterwards, write down one thing they told you that is still to come, with its date, and message them about it when the date arrives. That is the second message, and it is the one that tells them, and you, that the friendship is back on.

None of this needs an app. A line in your notes with their name and one dated detail will do. What does not work is a warm catch-up followed by “we should do this more often”, because that sentence is where most reconnections end.