Corporate entrepreneurship is an art. You have to learn to win for the team and for the company, to break just the right number of rules, and to swing almost for the fences.
Train like you fight: when deployed to Kuwait, my platoon of violated the safety guidelines during training exercises to master breach operations. Instead of traveling in our Bradley Fighting Vehicles up to a breach at 10 miles an hour, we would travel at maximum speed in wedge formation until the last minute, when we would form a line, with the trail vehicles slowing only enough to let the lead vehicles advance. Once in line, we would again travel at maximum speed through the breach using infrared cameras to see through the smoke and keep appropriate distance from the vehicle in front. My view was that it was actually more dangerous to train at 10 miles an hour because people would not be able to operate at combat speed during an actual engagement.
Think about your billionth user, build for your first customer: In 1999, I started working on a horribly named, but forward-looking social networking site: azazoo. I left my job as a patent attorney and taught myself Drumbeat + Access, then ASP + SQL. I wrote over a hundred pages of patent applications and created over a hundred pages of code. I was focused on the problems that people would have when everything was tied into a single identity, authorization, and identification system, from banking to building access to little league baseball management. For my efforts, I received a patent that was later sold via the Ocean Tomo auction. What I did not receive was a customer base. I was so focused on problems that people were going to face once my system achieved universal adoption that I failed to make my offering interesting to anybody. Facebook came along a few years later and in the matter of a year had achieved the scale that I had dreamt of by creating a valuable application for its first set of customers: students at Harvard. Mark Zuckerburg probably did have some ideas about how the world would be when everybody was on Facebook, but he focused on making his first customer happy and worried about the future later.
People want simplicity: when I was working as a developer for the Marine Corps Institute in 2001, there was an initiative to transition to electronic testing. A project had been started several years (and over $10 million) earlier, and the product of this investment was a couple of exams that could be taken at a handful of locations and less than ten percent adoption. I pitched the idea of enabling people to submit standard bubble-sheet exams using a simple web-form: build it once, and it works for all of the Marine Corps Institute's 100+ exams. Once we received the green light, we built the initial prototype over the course of a month. Despite, and perhaps because of, its simplicity, the application was very popular, particularly for the Marines deployed overseas who often had their exams delayed or lost in transit back to the United States. Within a year, most of our exams were being submitted electronically. A couple of months ago, I learned that the Marine Corps Institute has discontinued paper exams entirely.
Create the bridge between a successful experiment and product team adoption: Speed Launch was an experiment I conducted to see how my employer evaluated experiments. Over the course of a few months, we built a prototype of a very basic application launcher. The application received positive press and was downloaded over 90,000 times (more than most of the formal prototypes), but the product teams did not know what to do with the idea. To grow the project from 90,000 to 900,000 downloads would have been a challenging but manageable task, but even with those numbers it is not clear whether any product team would have found it interesting. It would be great to create a mechanism for select experiments that are not adopted by a product team to be placed in an affiliate "proving grounds" company. If a company does not invest in small yet promising ideas in hopes of growing them into meaningful business opportunities, they should stop incubating entirely.