The alphaWorks Group (on Facebook)
Thanks to David Gee for adding me to the alphaWorks group on Facebook. I am honored to be a member. I searched through the past 17 years of my blog and found a lot of references to aW. Perhaps the most meaningful is one from 2006 when I had attended the tenth anniversary of alphaWorks. I added to the post a section from my book, Net Attitude, where I wrote about the origin and vision shared by the team. The post is here. Andrew Morbitzer should be added to the group. He is the one that made the aW website happen with no budget. Elan Freydenson did the Domino coding and should also be invited. As soon as I post this, I will invite them both. At another time, I will post where the name alphaWorks came from and my conversation with Lou Gerstner about it. The following paragrpahs are the post from 2006.
There have been several stories here about alphaWorks. Today is a special day as IBM celebrates the tenth anniversary of the program. It was an honor for me to be part of the event in San Francisco. I made some remarks today about why and how alphaWorks was created but I decided to go further here and republish part of a chapter of my book, Net Attitude, where I gave some background on what alphaWorks is all about. It was part of a bigger subject called “Organizing to get things done”. Today we might call it collaborative innovation.
From Net Attitude (Perseus Publishing), November 2001
The most important ingredients to accomplishing great things as an e-business are to find, attract, recruit, hire, motivate, and retain really great people. Every year the crop of students gets better so you have to continually raise the bar — look at every movement of staff and ask yourself if you are improving your hand. Everyone has to not only bring something to the table but bring unique value to the overall equation. When things are working right the whole organization breeds and feeds on itself. If the caliber of your team is high, there’s a much greater likelihood of being able to attract additional high caliber people. Once you have them it is critical to nurture and support Net Attitude and to have creative programs to take advantage of their skills.
Every CEO I spoke to during the 1990’2 wanted to know how to make e-business web projects go faster. Every CIO I have met worries about e-business web projects going too fast. The CIO has spent decades getting information technology under control and making it reliable. Fast moving projects are sometimes in conflict with that goal. The solution to the dilemma is multifaceted but one key element is to have a “Skunk Works” where rapid prototyping is the modus operandi.
The Skunk Works
As far as I can tell, the origin of the term Skunk Works was at the Lockheed Corporation. For over a half century, the Skunk Works built a reputation that is unique in the world. Almost routinely, this elite group has created breakthrough technologies and landmark aircraft that redefined the possibilities of flight.
The Skunk Works was created to design and develop the P-80 Shooting Star, America’s first production jet aircraft. Since then they have created a string of firsts. In the 1950’s was the U-2, which to this day defines the possibilities of high-altitude jet aircraft. Then there was the SR-71 Blackbird which, with its titanium airframe is still the fastest jet aircraft in the world. The F-117A Stealth Fighter, which incorporated low-observable technology into an operational attack aircraft, created a revolution in military warfare. Its capabilities were demonstrated dramatically in combat during the Gulf War.
The company, now Lockheed Martin, says the key has been to “identify the best individual talents in aviation, blend and equip them with every tool needed, then provide complete creative freedom so they may arrive at an optimum solution in short order.” This simple formula is highly effective not only for creating state of the art aviation but also for any kind of corporate endeavor.
Lockheed Martin Skunk Works continues to serve as a wellspring of innovation for the entire organization and as they build advanced aerospace prototypes, and contribute to technology research and systems development. Lockheed Martin says this happens because they are “not big on titles or protocol – just getting the job done, regularly meeting schedules on time and under budget.”
The Skunk Works got its name from the “Skonk Works” of Al Capp’s L’il Abner comic strip, where they had a hidden still in a secluded hollow. The name still fits, because exciting things continue to “brew” there.
Small teams with maximum freedom of action
Product development is typically managed in a very structured organization with multiple levels of management and a lot of controls. This can be effective in many cases and is probably necessary for extraordinarily complex projects like putting a man on the moon but this approach will likely not bring any breakthroughs. The Skunk Works uses a different model.
Small teams with maximum freedom of action, very flat management structure, and minimal controls can lead to breakthrough ideas – if the people are allowed to work below the radar tracking level of the larger bureaucracy.
(Small teams of really top people are also more productive, and have more fun, than a significantly larger team.) Skunk Works are also good at figuring out what key problems there are in existing systems — because the Skunk Works members have no vested interest in the success or failure of those systems. They can often solve problems that the larger organization can’t solve because the larger organization is too close to the origins of the problem. It is usually best to let the Skunk Works figure out what things they should work on as opposed to “assigning” problems or projects to them. Problems the organization thinks are most important may not be optimal ones for the Skunk Works to invest in.
The formal requirements processes typically used to determine what should be developed don’t always anticipate some of the most profound issues and problems. The Skunk Works often just stumbles into profound things if you trust them and give them freedom of action. The instant messaging system being used by over 200,000 people at IBM did not come about because anybody asked for it or because a strategic planning or requirements process called it out. A few Internet software engineers stumbled into it, tried it out, built a prototype, and then nurtured it. In a couple of years it became an indispensable application for the company.
A subtle but critically important element in a successful Skunk Works is executive support, or “air cover”. There needs to be a well respected and highly placed executive who trusts the “lunatics” who are out on the edge.
At times the executive will be scared to death that a project the Skunk Works is pursuing will fail, but has to have the nerve to place a bet on it and trust the team to come through. Visiting the team late at night or on a weekend, bringing pizza and soda, showing that he or she cares and has a clue about what the team is working on, even if they don’t really understand the details, are critical ingredients. The little touches motivate the team beyond belief.
One of the biggest challenges with a Skunk Works is figuring out how to take the prototypes developed by a small team with a “just enough is good enough” mentality and integrate it with a more disciplined development process of the larger organization. In effect you have a tiny gear spinning at high speed trying to synchronize with a much larger and slower turning gear. One approach to solving this dilemma is to use an “impedance” matcher. Think of it as placing a third in-between-sized gear between the small one and the large one. Rather than a gear, of course, it is a small group of people whose mission is to adapt the prototype to the standards of the larger organization. Their focus is not developing it but rather adapting it, smoothing over the rough edges, and getting it into good enough condition that the larger organization will look at it and say it is good enough to be adopted and taken to market or put into production. The result is a speed to market that is a little slower than pure prototype but much faster than the full-blown process. Without the impedance matcher the larger organization is more likely to view the prototype as a virus and seek to eradicate it.
Fail and fail often
A successful organization has to be willing to have projects that are going to fail. A process designed to keep failures from happening is antithetical to a Net Attitude for innovation. But you need to be able to declare a failure, move on, and not punish the participants for being assigned to (or even creating) the failure. A good process encourages people to submit ideas into the mill as quickly and as often as possible and allow others downstream to figure out which ideas are worth pursuing further. There should be no penalty for putting in an idea that gets rated “close – no action required”.
Skunk Works are a vehicle for developing new things, or for bringing alternative ideas forward – they are not a universal answer to all problems of innovation. For Skunk Works to succeed, the company at some point may have to be cannibalistic. Children that come to life through the Skunk Works have to be able to eat their parents. In many companies there are countless examples of brilliant ideas and technologies that came to life in Skunk Works fashion but were then squashed by the mainstream part of the company. These same innovations are often successful when they are brought to market outside of the company.
There are a number of Skunk Works scattered around IBM Research laboratories and other parts of the company. One thing they have in common is the challenge of finding a path to market for some of their ideas which have no clear destiny. During an early 1996 visit to one IBM’s Research laboratories, Irving Wladawsky-Berger, then general manager of IBM’s Internet Division, noticed a particularly interesting streaming audio technology that had potential use for Internet applications. In fact he saw numerous technologies that seemed to have potential. While the technologies were quite impressive, there was no clear “business case” to take them to the market. Irving asked me to figure out how to reinvent the process of getting these research-phase (often referred to as “alpha”) technologies out of the lab and into the market. I thought about it all that weekend and then it hit me like a ton of bricks — all the bells and whistles went off in my head. Put these technologies on a website and offer them as free downloads and let the market tell us what they think the technologies are good for. We could put basic legal protections in place, create an easy mechanism for feedback, and perhaps even build a community around these early stage technologies. Without much planning, reviews, or analysis, we decided to implement alphaWorks quickly.
There were a lot of questions. Aren’t we giving up control of the technology? How about if someone takes our idea and turns it into a big business? By putting great technologies on the Internet for anyone to download, are we giving up intellectual property and will we later regret it? Yes, it is surely giving up some control but we believed there was more upside than downside. The technology candidates for alphaWorks were in some sense “orphans”. If the business case was clear for them they would be adopted by a product line of business and would be developed into products.
These orphans seemed to be brilliant ideas but the application for them was not clear. If a lot of people download them and find them useful we will get feedback on what they found them useful for. This could help us go the next step toward product development. If nobody downloads them or the feedback is negative we could kill the project and redeploy the resources to other more fruitful areas.
There was no formal organization; it would just evolve. That is how most important ideas flourish. If there is conviction in the idea, just do it, don’t’ study it. Don’t focus on who reports to whom. Just focus on getting something into the market and then let the market tell you what is good and what isn’t. If the idea takes hold you can build a more formal organization later — organization can kill an entrepreneurial idea if it is formalized too early.
A full-blown web site for alphaWorks was built in weeks. A couple of college interns who didn’t know things like this were supposed to be hard created a very impressive site. The legal team created a very simple agreement that said that if someone downloaded our technology they could do whatever they wanted to with it – except sell it. A process was put in place to enable IBM researchers to introduce one of their technologies onto alphaWorks.
Executive “air cover” was in place and alphaWorks came to life – downloads started happening and feedback started to pour in. Outside-in.
alphaWorks morphed from a site for “cool orphan technologies” to an effective way to surface emerging technologies and create paths to market for them. A community of hundreds of thousands of early adopters, entrepreneurs and innovators emerged that provided “headlights” to enable the company to see how people are thinking about the technologies, what challenges they face, and what features and support they would like to have.
As a byproduct of reaching out and forming a community the company received positive press coverage and was able to build mind share about its technology.
Ten years later alphaWorks is thriving and innovating and reinventing itself.