Illustration by Angelika Manhart
Lab Confidential

Startup Science: a How-to Guide

By Rachel Denison

Just got your first Assistant Professor job? Congratulations! You’ve spent the last goodness-knows-how-many years immersing yourself in your specialty, mastering cutting-edge techniques, and honing your benchcraft. But few of these skills can help you now. To meet the challenge of running a lab – of managing people instead of data – get some tips from other scrappy newbies. No, not your grad school nemesis who’s starting the third year of a job that’s better than yours. Why not look beyond academia altogether, to the scrappiest kid on the playground: the startup?

New companies have a lot in common with new labs. Both are filled with super-smart folks motivated by personal achievement and a desire to change the world. Both are trying to create an innovative product, be it physical or intellectual. And both are trying to secure a first round of funding. But in terms of keeping their people happy, motivated, and working crazy hours instead of spending half the morning reading PHD Comics, startups can teach research groups a few things:

Stand up. Many startups have daily “stand-ups” (meetings conducted standing up, to keep things brief), in which every team member gives an update on recent activities, immediate plans, and roadblocks that could benefit from the group’s input. Daily stand-ups might be too frequent for the more painstaking pace of lab work, but by devoting the first half hour of your weekly lab meeting to a stand-up, you’ll provide a social incentive for weekly progress and build community around shared problem-solving. Check out a report by Michael Hicks and Jeffrey Foster from the University of Maryland, complete with survey data from their students, to see how they successfully incorporated stand-ups into managing their research groups [1]. No more not knowing what the guy sitting at the weirdly small desk has been doing for the past two months (or years).

Carve out non-project time. At your individual meetings with group members, include some “non-project time.” Encourage your lab members to use these chats to discuss things that aren’t science per se, but are nevertheless intrinsic to their work. Topics might include troubles with motivation, life events intersecting with work, and career goals and questions. Making non-project time routine will help you get to know your lab members and counteract a common rule of thumb among grad students: that the PI should think of them as invincible mini-gods of the bench. Keeping non-science lines of communication open benefits everyone when the deities of life outside the lab start to have their way (reliably, in year three of the Ph.D.).

Give and receive regular feedback. How often did you get structured feedback about your overall progress and development as a scientist during your Ph.D. or postdoc? Maybe once a year at a thesis committee meeting? Maybe never? How often did your advisors solicit your feedback about their performance? (Rhetorical question.) Instituting regular bidirectional evaluations, which are standard at many companies, will let lab members know how they’re doing, help you grow as a mentor, and reduce the terrible burden of uncertainty for everyone.

Get better at communicating. To make the previous suggestions effective – to listen to, respond to, and critique your group members empathetically and clearly – you might have to improve your communication skills. Companies have management training, and so should you. Attend a workshop, read a book, take an online quiz and figure out how to deal with the results.

Define short-term goals and rewards. Startups move fast, and goals are often clearly defined in the short-term: make this product component; hit that sales number; increase website traffic by 15%. In science, the rewards that matter are much less frequent: you’ll publish a first-author paper once a year if things are going well. Do everything you can to define shorter-timescale goals and rewards with your lab members. Work towards conference abstract deadlines, and celebrate when the abstracts go in. Ask lab members to present periodically in lab meeting, and praise them for a job well done. Delight in a small result or solved problem before discussing how many more steps there are to go. Evaluations and stand-ups also serve this aim by providing frequent and predictable social rewards, which are essential for bridging the long and uncertainty-filled gaps between great results or big publications.

Build community. The most successful startups, and often the best labs, feel like a unified force. To emulate the highly interconnected project teams typical of startups, make an effort for every lab member to contribute to at least two different projects, and for at least one of these to be collaborative. Pairing new grad students with a postdoc or senior graduate student is a natural way to transmit lab knowledge while creating internal group ties. Cement those scientific bonds with regular social events or by taking a lab trip with plenty of just-for-fun time. Finally, write a mission statement for your lab, defining its community values. Whether or not you ever show this document to anyone, the ideas it contains will likely, through your words and actions, come to shape the culture of the lab.

Science is not a normal product, of course, and research is more like wandering around in the woods than assembling a log cabin. Not all startup strategies will translate well to the “lab space” (as the Silicon kids would call it). But if we want our lab members to do their best work and do it consistently and with pleasure, don’t we owe it to them to learn what we can from a domain that has management and team building down to more of a, well, science?

[1]  Hicks, M. and Foster, J.S. (2010) Adapting scrum to managing a research group. Department of Computer Science Technical Report #CS-TR-4966.


Rachel Denison is a Postdoctoral Fellow in the Department of Psychology and Center for Neural Science at New York University.