The first meeting goes well. There is real scientific overlap, both sides are interested, and you leave agreeing to work together.
Then eighteen months pass, a few emails are exchanged, and nothing exists.
This is the most common failure mode in international collaboration, and it is not caused by lack of goodwill on either side.
Why promising starts stall
Four reasons, and they compound:
No first concrete task. "We should work together" is not something anyone can start on Monday. Without a specific first deliverable, the collaboration exists only as an intention, and intentions lose to deadlines every time.
Waiting for funding. Both sides agree to collaborate once a grant comes through. The grant takes a year to write, six months to be decided, and often fails — and in the meantime the relationship cools.
Each side waiting for the other. Politeness on one side and uncertainty on the other produce a stalemate where both think the next move belongs to the other person.
Nobody owns it. Both principal investigators are busy; without someone whose actual job includes moving this forward, it stays at the bottom of two lists.
The small first project
The single most useful thing at this stage is a small joint output that can be finished in three to six months with resources both sides already have.
Candidates that work:
A joint review or position paper. Needs no new data, no ethics approval, no equipment — and it forces the two sides to align their language and find out whether they actually agree about the field.
A re-analysis of existing data from one side, with methods from the other.
A small pilot study at a scale that does not need external funding.
A joint proposal itself — writing one together is a real collaborative task, and even an unsuccessful one leaves you with a shared plan.
What this buys you is more valuable than the output: you learn how the other side works — how fast they reply, how they handle disagreement, whether commitments hold — while the stakes are still low. Finding that out during a five-year funded project is much more expensive.
What to agree in the first three months
Five things, none of which requires a formal agreement:
The first output and its deadline. Specific and near.
Who does what. Written down, even informally.
Authorship on that first output. Easy now, hard later.
A regular call — monthly is enough, but it must be in the calendar. The habit matters more than the content.
One person on each side who owns the collaboration. Often the best choice is a postdoc or senior doctoral student rather than the two principal investigators — they have the time, and it is a substantial opportunity for them.
The asymmetry to be aware of
In most partnerships between a well-resourced group and a group with less, the smaller side carries more risk in the early phase: more of their limited time, less certainty about whether it will lead anywhere, and less ability to absorb a failure.
Three practical consequences for the better-resourced side:
Be explicit about what you can and cannot commit. Vague optimism about future funding is costly for someone planning around it.
Make the early work count for them. Ensure the first output is publishable and that their people are visibly on it.
Do not let the pilot phase run long. A series of small unfunded collaborations that never becomes a funded project is, over time, a transfer of effort in one direction.
Four signals it is not going to work
Worth naming, because the alternative is a slow fade that neither side ever declares:
Repeated missed commitments on both sides over six months.
Scientific interest that turns out to be shallower than it looked — sometimes an apparent overlap is two different questions using the same words.
No response to a direct proposal. If you send a concrete plan with a deadline and hear nothing after two follow-ups, that is an answer.
Only one side ever initiates. Over a year, this pattern is reliable information.
Ending it cleanly is better than letting it drift: a short, warm message saying the timing is not working and leaving the door open preserves the relationship for later. Silence does not.
When the first output lands
Three things worth doing immediately, while there is momentum:
Agree the second project before the first is fully finished.
Use the first output as evidence in a funding application — a joint publication is the strongest possible demonstration that the collaboration is real.
Expand the circle: bring in a student on each side, so the collaboration does not depend on two people continuing to be available and interested.
That last point is what turns a personal connection into an institutional one — and it is the difference between a collaboration that survives someone changing jobs and one that does not.
Why do promising first contacts stall?
No concrete first task, waiting for funding before starting, each side waiting for the other to move, and nobody whose job includes moving it forward.
What kind of first project works best?
A small joint output finishable in three to six months with existing resources — a joint review, a re-analysis, a small pilot, or a proposal written together.
Who should own the collaboration day to day?
Often a postdoc or senior doctoral student rather than the two principal investigators — they have the time, and it is a substantial opportunity for them.
What signals that it will not work?
Repeated missed commitments over six months, shallower scientific overlap than it appeared, no response to a concrete proposal, and only one side ever initiating.
