Late last year I was invited to present some of my thinking on Social Media at an external conference. It was my first communications meeting and a great opportunity to run my thoughts, and suggested internal social media guidance, by my peers in the pharmaceutical space.
I arrived at the meeting, dropped my stuff off, and proceeded to rustle up a cup of coffee before the first keynote.
And this is where it got a little embarrassing.
They had a coffee machine. It was one of the fancy ‘European’ style pod machines, with levers, knobs, buttons, and a lot of different flavour suggestions, but no clear instruction as to how to get deliciously dark caffeinated goodness into a cup. I felt like I was standing there for hours, feeling confused and bewildered, before someone tapped me on the shoulder, and kindly offered to help.
Delicious cup of caffeinated goodness in hand, I bumbled through heartfelt thanks and got talking to my new best friend. Quickly, and somewhat surprisingly, the conversation turned to complex adaptive systems and the merits and pitfalls of agent-based modeling strategies.
Hang on a second; was I at the right conference? This would have made perfect sense at the conferences I used to attend, but not what I was expecting at my first communications meeting. Was I in the right place?
As she told me her name, it dawned on me that I knew of the book she’d written a couple of years ago (Joystick Nation). At the time JC was consulting for a company called Kaggle, and that she thought I should speak with the founders as she felt there’d be some good overlap between what we do, as a research driven pharmaceutical company, and what they do (host data science competitions for people asking questions of data). Following an introductory email and a brief teleconference, I began to think more about how this all could be used to our advantage.
So, over the holiday break, and in collaboration with two colleagues from Research, I structured a data set that would be amenable to this kind of a ‘crowd-sourced’ competition.
I then pitched the idea to the U.S. CIO and head of Global Innovation Management who agreed to provide the prize purse of $20K.
After shepherding this through our internal review processes, as scheduled, on Friday March 16th, the ‘Predicting a Biological Response’ competition went live.
By the end of the first day, without any formal announcement, we had people ‘playing’ – and submitting solutions to a problem we’d posed, and that is of real pharmaceutical interest. Today, at the time of this writing, and not even a month into the competition we have 208 teams, comprised of 260 players, who’ve submitted 1409 solutions; 89 of which are better than our best initial benchmark.
The official press release can be found here and a nice write up by PSFK can be found here.
My longest post on tumblr, but I thought I’d share, because I really love how I’ve turned being an incompetent with a fancy coffee machine, into a win for my organization.