Past As Prologue: The Tale of The Binge-o-Matic
(Note: This post was originally published here: http://blogs.law.harvard.edu/cesarbrea/stories/storyReader$8 on November 26, 2003)
I left Bain in the summer of 1999 to join ArsDigita, a small firm of ultra-talented software engineers led by Philip Greenspun, and famous for its open-source application framework, the ArsDigita Community System. Philip's ideas for online communities as the best use of the Web, and of open-source software as a more effective means of realizing such constructs appealed to me on many levels, and I felt I could make a real contribution to the firm. So I signed on as the second "business guy".
While, the rest of The ArsDigita Story has already been told from many different perspectives, this little fragment has stuck with me (with apologies to friends who have heard me tell it one time too many). It's said that "within every insanity lies a grain of genius," and for me this proves the point.
Back then, ArsDigita World Headquarters was located in half of a three-story duplex house on Franklin Street in Cambridge, leased to us by Elsa Dorfman, the portrait photographer. (In form, Elsa had leased the space to Tracy Adams, an ArsDigita co-founder, who constantly had twenty or so "friends" over for a round-the-clock "programming party", so that in practice we could run the firm without violating zoning laws.) One Friday morning, an entrepreneur -- let's call him Pete -- washed up on our doorstep. "I've got $2 million in venture capital," he said, "and I want to build the killer site for college kids who want to party."
In those days, we sometimes conducted business in the living room, where we demoed on the 45" Gateway monitor from a wireless keyboard, while our guests sat on a well-worn chintz sofa next to a Steinway baby grand. But more often, the real conversations happened later, out of the office on the hoof. On this particular day the four of us (Pete, me, Philip, and of course Alex, Philip's Samoyed) set out across the Charles River for the Harvard Business School campus. Philip felt the setting might discipline our talk and make a convenient pit stop for Alex.
We strolled along, heads down and hands behind our backs, like diplomats. Philip asked the obvious question: "what makes you think anyone will be interested in coming to your site?"
Pete explained that he had thought up the killer application -- the "Binge-O-Matic."
We sat on a bench in one of the quadrangles, under some trees; I thought I heard whispers in the wind rustling their leaves ("Don't be an idiot!" "Remember profitability!"). Pete went on: "So, the idea is that college guys come to our site, and enter in the number of times they've been sick from binge drinking in the last week. Then they see where they stand on a graph of their peers' answers." Oh. "The best part is, I serve them up ads based on where they are on the curve. Sober guys who need to loosen up get coupons for Captain Morgan's. Party animals get invitations to AA!"
Physically I was still at HBS. But I wasn't sure out-of-body me was in Kansas anymore. Philip smiled at Pete and turned to me. "So Cesar, you're our business guy. What do you think?"
"Well," I started in soberly, "I have three concerns."
"First, I'm not sure I can see this sandwiched between GE and GM on our S-1 statement when we file to go public."
"Second, I can already see the lawsuits from distraught parents of MIT freshmen who die trying to get 6-sigma out on the curve."
"And third, I can't see going home to my wife and kids and saying, 'Guess what Daddy built today!'"
Pete looked deflated. Philip said he thought the idea was interesting from an engineering perspective. But I had brought it strong, so conventional thinking prevailed. For a number of months after that, I got emails from Pete's company -- some memorably graphic -- until one day the messages started hawking banners, mugs, and sweatshirts, and then stopped coming altogether.
Several months later, I picked up yet another magazine blowing kisses at Jack Welch. Among the tales of many coups GE had pulled off was a story about Jeff Imelt's work at GE Medical Systems, which sells CAT scanners. Someone at GE got the bright idea to plug them into phone jacks, so they could be monitored and re-calibrated remotely instead of sending technicians out to do it. Another smart GE person figured out that since they were collecting performance data on CAT scanners, they could benchmark how effectively different institutions were using these expensive machines, and sell consulting services back to the laggards. Jeff Imelt is now CEO of GE. (So maybe Friendster users are right: it is all about the hook-up...)
Around the same time, we landed a contract to build a web application for Cyrano Sciences Corporation. Cyrano Sciences had developed a handheld digital gas chromatograph, aka an artificial nose (the Cyranose 320 was featured in the December 1999 issue of Wired Magazine). Somehow it occurred to someone that Amazon-style collaborative filtering for smell might make an interesting way to apply the Cyranose. The idea was to sell the device to wine lovers, who when they tasted a wine would also "sniff" it with the Cyranose to capture a very precise "noseprint" for each wine, then upload those noseprints into our web application to support recommendations: "this wine's print matches that one's; I liked this one, therefore I might like that one too." Interesting. But it turned out the money-making application for the Cyranose wasn't B2C, but B2B: process control in petrochemical plants. Cyranoses installed in Southern California refineries would sniff-n'-share, and our web application would send alerts to operators saying things like, "Gee, this plant has begun to smell like a couple of others that blew up last month, maybe you should open some windows and call the PR department to handle the neighbors' complaints."
The insight of course is that these are all essentially the same application. Our business was based in part on building innovative solutions for clients and then rolling generalized versions of those innovations into our open-source toolkit, which would then be the starting point for future innovations. So it dawned that having a general framework to help envision and anticipate useful innovations for web-based collaboration solutions would be a useful thing.
Boiled down, that framework consists of three questions: What could people share that would be of value to each other? With whom (within what natural affinity groups) will people be motivated to share? And, how can we structure the sharing to minimize the effort to contribute and maximize the consumability of the shared thing?
Newspapers that put up discussion boards as adjuncts to news articles have very poor answers to these questions, so if there's any content in those boards it's generally pretty bad. Random rants from unqualified contributors generally have no value. There is no clear affinity among the posters, other than the momentary interest in the same article. And a blank text field means for content to be good, the author must have the time and talent to write clearly, and the reader must have the time to sift through posts to get to the good stuff. For the same reasons, I'm a little dubious about a newer experiment, theSpoke.
Which brings me to 2003's social networking solutions. Relationships are valuable to share. People are comfortable sharing them in the right, private context: circles of friends, certain groups of co-workers. And some of them make it really easy (in some cases, automatic) to contribute and consume contacts. Of course I'd like to be able to say that I saw all of this in the framework I described above, but I jumped into Contact Networks first and it only dawned on me later. Maybe (hopefully) better lucky than good.
But, should anyone lay claim to be the next Internet visionary based on a solution-du-jour, the lesson of the Binge-o-matic is that we haven't even started to get creative!