Over the years I've been involved in quite a few collaborations of various sorts. Many have turned out to be productive. Some have just fizzled. Others have been a royal pain in the arse. And yet others... well, it's too early to tell.
Thing is, nobody really tells you how to go about setting up a collaboration to maximize its chances of success. I don't mean the whole "identify a putative collaborator and contact them" process, more the "okay, we've agreed to collaborate, now what?" side of things. And yet that's often the most crucial part of it all.
Obviously you want to set things up in a way that benefits both you and your collaborator. But it can be hard sometimes. If your collaborator is more junior you don't want to stomp all over them,* and if they're more senior/prominent sometimes it can be a bit daunting to propose something where you're equals or with you, the more junior person, with the larger/leadership role.**
Interestingly some of the most dysfunctional collaborations I've seen have been between peers. It's not always clear to me why. May be something to do with both participants being under the same pressures.
No matter the circumstances, you've got to do what's best for the collaboration.
And that's what counts in these things - the collaboration.
So right at the start you need to have what can sometimes be a difficult conversation. Or conversations. Before anything really gets started. Don't wait to do this.
Here's what you need to thrash out in advance:
1) Who is going to do what.
Okay, it might seem obvious who is going to do what, right? After all, you've likely set up the collaboration because you need expertise or a system you don't have. But it never hurts to be very clear about the scope of the collaboration and what the expectations are for each participant. Also talk about what personnel in each lab are going to be assigned to do the work. Don't take it for granted that your collaborator has someone who can start on things immediately.
When you're hoping to get things done by. Everyone recognizes that science never goes as planned and timelines are at best wild guesses, but deadlines, even squishy ones, help get things done.
This is a biggie. Many a collaboration has been derailed at the publishing stage. And the authorship battles can be nasty indeed. Try and head them off by settling on some guidelines for first and last (senior/corresponding) authorships. And everything in between. Obviously you'll need to revisit things with each publication, but it helps if you start with some broad guidelines.
4) Status on grant proposals.
In my experience most collaborations are started with a view to obtaining funding. Who is going to be PI? Co-I? Will it be a multi-PI proposal? Or are you thinking sub-contract? Obviously that can all change as the collaboration evolves, but again, starting with a general idea of how it's going to work can prevent a whole world of hurt down the road.
There's probably more, but in my mind those are the big four. Finally, if you have that conversation in person or by phone, send a follow up email outlining everything you agreed on. Six months or a year on you don't want to have a disagreement over what you agreed on. Create and keep a (e)paper-trail.
A good collaboration is a wonderful thing. When you're embarking on one, make sure you do the legwork necessary to give it a chance to be successful.
* Unless you're a dick.
** Pretty, pretty please Herr Professor Biggus Wiggus, if it's okay with you, and I'm alright if it's not, but it would help me a lot if I were, let me be PI? Maybe? Perhaps?