Based on my recent question on simultaneous working papers strategy and relevant insights by Stephan Kolassa, I would like to solicit the community's thoughts, advice and best practices on the topic of finding potential collaborators and co-authors (especially from a beginner's perspective).

Initially, I thought that it is a rather simple question with a most likely and obvious answer to it is simply to intensively network, interact with relevant people and publicize your own ideas, thoughts and projects (in other words, build your academic "brand"). However, while the core idea is most likely to follow the above-mentioned common sense, on second thought, there might be much more to it, such as various important strategic and tactical considerations that I could have missed.

  • @ff524: While the question you reference is closely related, I think that it would be beneficial to leave my question open, because: 1) the former one has been asked a while ago and doesn't seem to be active since then; 2) the existing two answers are limited in coverage and, thus, the answers domain could be improved. – Aleksandr Blekh Sep 10 '15 at 7:59
  • 1
    If the question is basically equivalent but the answers on it are lacking, the appropriate procedure is to offer a bounty on the old question to draw more attention to it, not to leave a duplicate open. – ff524 Sep 10 '15 at 8:01
  • @ff524: IMHO offering bounty might not always help in drawing more attention to a question, especially if it is suffering from deficient formulation. I humbly believe that my question is clearer and, thus, better in this regard. – Aleksandr Blekh Sep 10 '15 at 8:12
  • 2
    @AleksandrBlekh What is the other question lacking? Especially I cannot imagine going beyond the JeffE's answer. When it comes to strategic or tactical - even if you do everything right, finding a collaborator has a lot of random components (interests, time, character alignment, ...) so like with every networking-like process, for every dozens (or hundreds?) of opportunities one a few are going to fire. – Piotr Migdal Sep 10 '15 at 8:51
  • 1
    @AleksandrBlekh The structured way is to apply to a team, and you will be given both a topic and collaborators. For the unstructured - JeffE has his own style, but I think that it is easy to extract the massage: talk to many people with similar problems/approaches, and a tiny fraction of them may turn in collaborators (but it's a multi-iteration process; similarly, usually you don't propose during the first date). – Piotr Migdal Sep 10 '15 at 9:43

Here are a few key observations I've made about the process of starting collaborations:

  1. Most people one would like to work with are already quite busy, and have many random people come to them looking for collaboration. If you want to be taken seriously, you need to have a specific and detailed idea of a way that their work can make a difference to your goals or your work can make a difference to their goals. Once you start discussing, you might end up somewhere totally different, but clear and specific thinking is key to getting things started: I suggest Heilmeier's Catechism as a useful framework for your thinking.
  2. Most collaborations thus start with a discussion with between PIs, but are only fully realized if a graduate student or postdoc becomes involved to do a lot of the work. Early career researchers may play the role of both PI and postdoc themselves. Be careful, however, to keep acting as a collaborator rather than turning yourself into a free postdoc.
  3. Almost no interactions turn into collaborations. A good rule of thumb is that for every 10 people you meet, 1 will become a potential collaboration; for every 10 potential collaborations, with 1 there will be meaningful follow-through on the possibility, for every 10 with meaningful follow-through, 1 will actually turn into a significant collaboration. Don't get discouraged, but also don't waste time pursuing people who indicate they aren't interested beyond the point you are at.

Not the answer you're looking for? Browse other questions tagged or ask your own question.