Most language exchange partnerships die in the second week. Not the first, because the first carries the novelty of meeting a stranger from elsewhere, and not the third, because by then the survivors have settled into a habit. The second week is when both parties have to schedule a call into a real week, with real other commitments, and one of them quietly decides it isn’t worth the friction. The other side notices, doesn’t ask, and the thing dissolves without anyone being rude about it.
This is true of nearly every exchange I have set up myself and nearly every one I have heard about. It is not a comment on the people involved. It is a comment on the structural fragility of the arrangement. Two strangers, neither paying the other, meeting voluntarily, often across time zones, to do something that requires more concentration than a normal conversation — there is almost nothing holding it together besides goodwill. Goodwill in week one is endless. Goodwill in week two is finite.
If you want to be in the small fraction of pairings that survive past month one, almost all the work happens before the first call. Most people pick a partner on the basis of the surface fit — same age, broadly compatible interests, the right native and target languages. This is too low a bar. The fit that matters is motivational. You are looking for someone who is trying to learn your language for a reason at least as serious as your reason for learning theirs. If you are preparing for a job that will require you to work in Mandarin and they are casually learning English because they enjoyed a series, the call will be lopsided, and the lopsidedness will get worse with each session until one of you stops replying.
The second filter, which most people are squeamish about applying, is age and life stage. A working adult and a college student rarely sustain an exchange beyond a few weeks, not because they have nothing to say to each other but because their availability looks completely different. One can call at 10 p.m. on a Tuesday; the other was asleep at 9:30 because they have a meeting at 7. The exchange that survives is the one where both parties can plausibly meet at the same time on the same day every week for a year. Anything that requires bespoke scheduling each week will collapse under the weight of its own logistics.
The third filter is purpose. Is the call going to be unstructured conversation, or is it going to be the more useful but more demanding format in which one of you presents something — a news story they read, a film they watched, a problem they want to talk through — and the other responds, asks questions, corrects, and then the roles reverse? Unstructured conversation is what beginners do because they cannot do anything else. By intermediate level it stops working. There are only so many times you can ask someone what they did on the weekend before the call becomes a polite waste of your respective hours. The exchanges that last past three months are almost always ones where both parties have committed to a structure that demands a little preparation. The preparation is what makes the call worth showing up to.
The first call itself ought to be administrative, not social. Agree on time. Agree on duration — fifty minutes is right, an hour is too long and feels like a meeting. Agree on language split: half and half is standard, but it matters that you both want it that way. Agree on correction style: some learners want every error flagged, others want a single corrected version at the end, and a mismatch here builds quiet resentment fast. Agree on what to do if one of you cancels: reschedule within seventy-two hours, or skip the week entirely. These small contracts feel excessive on call one. By call ten you will be grateful for every one of them.
The thing nobody tells you is that the right partner is usually not the most charming one you find. The right partner is the one who shows up on time, who has done a small amount of homework, who corrects your obvious mistakes without making a meal of it, and who treats the arrangement with the unromantic seriousness of two people doing a job together. The charming ones are wonderful for the first three weeks. The dependable ones are still there at month six, by which time you can both have conversations that would have been beyond you when you started.
A last point, because it always comes up. A tandem partnership is not a substitute for a tutor, and treating it as one will frustrate everyone involved. A tutor is paid, has obligations, can structure your learning, can correct you in domains you do not yet recognise as wrong. A tandem partner is a peer, a fellow learner, and what they can give you that a tutor cannot is sustained exposure to a person using the language for their own reasons. The two practices serve different ends, and the people who do best at intermediate level usually do both.
