I am going to take the perspective of someone trying to build a business based on novel science (as opposed to the effects of science on a more, shall we say, mainstream business). This is hugely important for emerging markets, where the attention to science is generally still minuscule, and where the bridges between academia and entrepreneur are almost completely absent.


Let me draw upon two of my recent experiences to illustrate the principles below. One venture is a machine-learning software company matching talent and opportunity around the world, Aspiring Minds; the other is Jana Care, a medical device company facilitating diagnoses for diabetes and helping patients and clinicians manage care. The former draws on recent advances in computer science, the latter on life sciences. Their similarities and differences are instructive.

The first step for a science-based entrepreneurial endeavour is to develop the science in a cluster where other scientists are located. Problem-solving is a lot easier in these milieus especially for a start-up that will have a limited number of scientists on board and will inevitably have to seek outside help. Having a good university or a cluster of established science-based companies around sure helps. For Jana Care, we do our science in Boston, a life sciences hub, though the company started

in Bengaluru. The algorithms company Aspiring Minds accesses talent in Boston and San Francisco, but it would be helped more if it were located there (it’s based out of Delhi and Beijing).

There is the idea of the death of distance, a term popularised two decades ago by Frances Cairncross (editor of The Economist) to describe our ever-increasing ability to communicate digitally (and therefore remotely). At the end of the day, though, I think nothing beats the convenience of proximity to fellow science-travellers for those working on cutting- edge science.

Second, where you do the science need not be the place where you implement it. I don’t mean to draw an overly stylised distinction between doing and implementing — as a colleague Eric von Hippel at MIT has frequently reminded us, “doing” and “using” feeds back strongly into R&D — but I’d encourage thinking about where the cost of experimenting with whatever you produce are lowest. That’s where you should try out the results of your science. Our medical devices are deployed across parts of South Asia before they’re deployed in Europe. An important caveat to experimenting efficiently, especially for health related sciences, is that patient safety and well-being are paramount — you can’t economise here.

Third, decide how much of the problem you want to solve. In Aspiring Minds, our algorithms to assess talent could, in principle, have been licensed to someone else to use. We need not have bothered with the rest: building a sales force, delivering millions of tests annually, etc. In practice, we had to build the whole ecosystem to demonstrate its value. For life

sciences though, it’s commonplace to do the science up to a point. Thereafter, one hands the results to an entity better suited to further develop the solution to the problem and is paid to do so. Think of research on finding new targets for drugs. A collection of scientists often characterise a target molecule, pass it on to a typically much larger pharm company, and get paid handsomely. Sometimes these drug targets pass through multiple hands. A downside to doing only part of the science is that the scientist entrepreneur doesn’t get to shape the destiny of their creation.

This principle is more general than in drug discovery; the question being, what part of the problem are you best suited to address? This requires cultivating an approach to partnering and being able to know when to stop working on a problem in favour of handing it off, and whether and how to redeploy resources.

Fourth, protect your intellectual property. Sometimes patents work, in some scientific fields and in some (geographic) jurisdictions. Otherwise, non-patent methods of protecting the technology are usually more effective — it needn’t be either or. For example, the tech in question could be protected by embedding it in a product, or a business system, making it less useful as a stand-alone imitation.

Fifth and most important, be aware of the limits of the science. They might come because of your inability to not only take care of other ambient constraints but also because of how good you really are compared to the world’s best. That requires humility, open-mindedness, and a willingness to learn. Often, sadly, this is the attribute most in short supply!