Translating Research

or why Academia does not create MVPs…

From a bioinformatics algorithm to a minimum viable product. It seems like a trivial process and should indeed be a no-brainer, however, creating real products from research work is almost an impossible task.

I’ve been doing computer science research applied to bioinformatics for over 5 years. I worked on a couple large-scale international projects and developed a few web applications. However, unless you are working for a large (EBI, EMBL, NCBI/NIH, SwissProt…) research institute, is very very hard to go beyond the “good enough” level and drive more user adoption.

And why?

The translational research prerogative is flawed. Creating real marketable products from scientific work implies an effort level that is beyond the reach of most research groups.

Let’s look at a simple scenario. You spend 1 year researching a new algorithm that improves the multiplication operation efficiency 10 times. You propose a hypothesis, you test it, you implement the algorithm as a web application that everyone can use, you publish the work in a high impact factor journal. What’s next?

Even if you can attract users to your new web application with your fantastic new algorithm, you do not have enough grounding to turn it into a product. At this stage, two things can happen: 1) funding ends, and you leave for the next big thing, or 2) you have new funding to research something entirely new. At this point, what happens to your state of the art multiplication algorithm application? It perishes in its purgatory without new features or updates… Perhaps will add an how to cite section so that everyone can boost your publishing index…

And herein lies the big problem. There is no funding to provide continuous support, to perform adequate user testing on all features, to create documentation and help sections, to add new features… Customer/user/researcher question and emails go directly to your inbox, where they will probably get lost in the confusion of your further projects. User testing, like A/B testing or more specific usability tests, are a mirage. You can’t pay users and you, since this isn’t a new social-network-game-something, can’t get enough potential users to try it out. Documentation and help? Well, you can put a FAQ somewhere, but that will have to do for the first release (and for eternity).

Adding new features or implementing significant improvements to the application leads us to the problems above: no time and no money. There are problem newer projects that require your full attention or, if it is something unique that you really want to continue developing, the funding will probably be nowhere to be found.

In sum, designing the algorithms and creating applications is just the initial challenge. Maintaining these new systems online, attracting users or adding new features is impossible to do. And these are essential requirements for a minimum viable product.

Pedro Lopes
Aveiro, Portugal