In science, as in all activities, sometimes you learn more than you are aware of while carrying them out. Keeping your eyes open as you go can be an incredibly valuable way to garner extra knowledge… as in this case example, where collaborating on a paper led to me pondering handovers, organisational risk, and how to strategically manage groups to mitigate that risk.
A bit of background about my field / experience for people outside university research groups
In university-based science, academics lead groups consisting of post-doctoral researchers (often shortened to post-docs), and students. The post-docs are staff who have a PhD, and therefore have at least three years’ experience in research (the shortest expected study duration I am aware of) — more if they have already held a post elsewhere. Students are usually making their first foray into the field.
Student projects are usually overseen by the group head and either another academic in a related field or a post-doc in their group. Whether it is officially recognised or not, a post-doc usually has day-to-day responsibility for managing the student and their project.
The post-docs are normally required to carry out projects of their own simultaneously (supervised by the academic, who is paying them for their efforts). The post-docs usually want some time for completely independent projects too, in the hope of attracting their own funding and making the leap to permanent employment as an academic.
Research projects aim towards original findings that are published in journal papers. These papers form the major basis of an early-career (finished student, or post-doc) researcher’s applications for positions, funding, and promotion.
The task that opened my eyes to this issue
My second ever paper was my first foray into being a collaborating author. Here, another student, Brett, was simulating a (nitrogen-vacancy) defect in diamond, trying to figure out which charge state (presence or absence of a single electron) was more stable under various conditions. This is still an important question for many applications, since the charge states have drastically different properties, and the wrong one can derail a finely tuned experiment or device.
The real lessons from this task came as Brett moved on to his next role. In particular around the handover of results and notes; although a pretty decent draft manuscript existed at the time, the rest of us had some questions about how the results had been treated… Unfortunately, Brett wasn’t around any more to ask. Sometimes, sporadic email contact just doesn’t have the bandwidth to convey all the subtlety you need for a detailed discussion. Luckily, he’d supplied us with the input and basic output files from his supercomputer work, and the scripts he’d written to process some of the data further.
With a bit of thought, a bit of reverse-engineering what he must have done given the theory used and the graphs generated, and several months’ time, we arrived at results we both had trust in (since Brett did good work), but also had the deep understanding of that only comes with cranking through the problem yourself. Of course, we were then much happier to modify the manuscript and run it past him again quickly for comment, and the paper was published shortly thereafter.
What I learned from this project (and reflecting on it and others since then)
This experience made me fundamentally aware of the risk that organisations (particularly scientific research groups) carry given the possibilities of staff changing roles, making career moves, or even requiring extended leave. Group aims can be set back by months if somebody has to pick up the cold pieces without any guidance — and the more specialised the knowledge base needed for the job, the worse off you are likely to be. It is very rare to find another staff member, for instance, with the precise background to pick up a materials simulation project without resorting to starting again from scratch.
There are two main aspects to this risk: the unforeseeable causes with immediate effect (e.g., death, significant injury, or resignation with minimal notice), and the longer-term causes (e.g., failure to renew contract, or career moves), which are often more amenable to managerial planning and strategy. In science, where post-doctoral researchers are almost ubiquitously on fixed-term contracts lasting at best three years — and those dependent on group leaders obtaining research grant monies — this second case is an ever-present part of group dynamics.
Managing the unforeseeable largely comes down to encouraging a culture of good note-taking, with regular and thorough communications up the management chain of activities — in university-based research, usually by weekly or fortnightly student-supervisor or staff-supervisor meetings as well as email. This at least gives new staff a thread of developments to date, and if done well, the reasoning behind them is also evident, along with references used along the way. The gold standard includes notes looking forward to future tasks, and the planning thereof.
A group-accessible backup server is also a good option if the group are disciplined about file structure layout, naming convention, and retaining only files which are both sufficient and necessary to understand the work. If the group can easily back their work up, and members can provide a key to understand how they index their files if they deviate from group norms (if such exist), then the barriers to another member taking over are lowered.
All of these approaches require some level of investment at the start of someone’s career in the group — depending on their level of prior experience. An incoming student will not likely have much expertise in keeping, e.g., lab journals or theory notes to a level accessible to anyone else outside their project. An experienced post-doc will hopefully have encountered the problem of interpreting notes at a later date (either their own, or someone else’s), and will have become more comprehensive in their writings.
A useful example to highlight the importance of the task in the initial meeting(s) is the relative time cost (e.g., an hour of planning at the start of the week, and an hour of documented reflection at the end) over the expected timeline (three years for a PhD), versus spending six months reinventing the wheel at the end when writing up… I’d rather spend the six weeks myself.
Either way, it behooves the group leader and/or senior staff to regularly inspect such records (I suggest at least quarterly), to manage this risk. It also mitigates any risk arising from an outgoing employee ignoring handover requirements and refusing to communicate. Handled properly, it can be an extraordinarily useful lesson in record-keeping — but it works best when such mentors can easily provide clear examples from their own shelves or hard drives.
The longer-term risks, particularly around contracts, are often poorly managed in science. This is usually due to the lack of funds available to advertise for, hire, and start new employees with any (let alone sufficient) overlap with the outgoing staff. It is also common that staff, knowing that the end of their contract approaches, apply for other jobs which then ask for start dates as soon as allowable given the required notice of termination — which makes planning the campaign to have replacement staff on hand nearly impossible.
The consequences of having a senior group post-doc leave can be devastating if not managed correctly. Research groups, which often operate on a shoestring budget, can rarely afford to double-up on skill sets. Senior students are often required to take up the slack in the meantime, but they cannot be expected to embody the range of abilities a post-doc has. They simply haven’t had the time in the profession yet. They are also usually about to leave the group themselves. Worse, what happens if the only student on that project has just begun? Or if there weren’t any students involved at all? The group faces operating, possibly trying to meet deadlines with external bodies (collaborators, funding agencies, industry groups) with a hole in the core skills they require.
Approaching this type of risk requires an ongoing strategy. In addition to running as described above, the group needs to be constantly training people in duplicate skills. Such training necessitates spending time and effort that could otherwise be exploiting post-docs’ current knowledge base(s) for extra outputs (in our case, primarily journal papers). Individuals may feel that such efforts are detrimental to their career prospects, since these are largely based on the number of their outputs. This feeling can be ameliorated by co-authorship on papers, and official listing as co-supervisor of students. On the outside, such a role can be spun as experience in training and mentoring, as well as project oversight. Explicitly exposing the post-doc to any planning and strategy activities (and thus, the skills) is another potential sweetener, if pitched in the right way.
A canny group which becomes known for the management training their alumni receive and can demonstrate is one whose alumni may well be sought after and fought over by other research groups — and also by the outside world. Attracting high-quality students and staff might become even easier if this reputation becomes well-publicised. Industry may be interested in sponsoring mutually relevant projects too, particularly if it gives them an edge in attracting the graduates afterwards… potentially allowing the leader to expand the group in ways that afford built-in redundancies and lessen the ongoing training load.
Of course, there is always a tension between keeping the documentation that most of us know we need, and actually finding the time to do so. However, a little bit of forethought about how to carry it out might just pay dividends — not just in my own work, but perhaps in my students’ too. If I’m ever lucky enough to run a larger team, I hope to expand the experiment. Of course, I’ll have to continue documenting my thoughts around it, and reflect on how it is succeeding and failing…