Skunk Works ★

At the end of college, I was interviewed for a job with Lockheed Aircraft Corporation, including their famed "Skunk Works"division. They had a huge skunk painted on the wall at the entrance to their work area. Everyone said the same thing to me: "We can't tell you what we do, but we sure have fun." I ultimately went to work elsewhere, but I occasionally wonder what it would have been like to work there. I don't know what work I would be doing, but I'm sure I would have fun doing it.

...organizations have the freedom to iterate and innovate early in the life cycle of their major products. As a project matures, the context becomes rigid, and innovation becomes "forced" and may appear in the guise of "innovation programs." While these programs are good at the divergent thinking component of innovation, they rarely do a good job of convergent thinking. The result is that novelty, valued for its own sake, finds its way into mainstream development where it incurs costs but leads to results that range from indifferent to disaster; "home runs" are rare. The net result is most often negative.

✥ ✥ ✥

A project must accommodate major innovations but must also keep an eye on risk. It is too risky to innovate too much in project development. Some projects have "innovation programs" that value divergent thinking. The fruits of these efforts often make their way into development. It is only rarely that an organization does an honest evaluation of whether such ideas actually added value; the value is often taken on faith. For example, the latest technologies are always held to have value in their own right; conversion to OO, or to components, or to patterns, is considered "good" without a second thought. Too often, these new ideas have either indifferent results or in fact increase cost. They may decrease time to market or decrease cost, but if they decrease cost at the expense of time to market, then the overall effect is disastrous if time to market is the highest business priority. And, in fact, any new idea can both increase time to market and cost in ways that may never be noticed, in part because of the stock taken in the buzzword value of the idea.

Yet projects become dead if there is no way to get paradigm shifts into the project now and then.

Therefore:

Allow a limited-cost SkunkWorks to form (as a SelfSelectingTeam) to develop an idea outside the constraints of project development, to build confidence in the idea. Give the SkunkWorks organization ownership and credit for the idea.

The organization is sustained by strong FireWalls that insulate it from the scrutiny of upper management and funders; in fact, the very existence of the SkunkWorks should be a secret. The idea is to keep the project off of management radar screens to foster the kind of innovation that leads to success before tradition and its constraints, as embodied in managers, can dampen innovation.

The success of the idea is assessed according to the fruits of the SkunkWorks effort: the ability of the resulting product to attract customers willing to invest time, money or people in building the product or in otherwise furthering the idea. The product must tangibly show positive results that differentiate it from the mainstream product line; if it is to thrive over existing external and internal competitors, it must demonstrate distinguishing market superiority. Directly moving new ideas into the business units rarely works. As a practical matter, this evaluation of success and the ensuing steps to act on it happen at unusual places in the management structure: at a higher level of management, in an organization that has venture funding, or by using the leverage that marketing can bring from customer needs statements and customer commitments. However, the technology's chance of long-term success is much higher if the skunkworks team includes developers who also have product responsibilities in existing products. They can become seeds for new development teams for the new products or, if they are very lucky, they can be conduits for introduction of the new technology into development organizations. Therefore, this pattern also depends on giving some small set of interested developers some limited amount of time to work with the SkunkWorks team in a GateKeeper capacity.

The SkunkWorks organization itself rarely can take a product all the way into production. It usually lacks the infrastructure, and sometimes the skill set, to build a solid product. This phenomenon is at the root of many well-known stories about large companies not being able to capitalize on their greatest inventions.

If the idea succeeds, the team should reap the benefits of the idea. The organization subsidizes some of the risk of the team under the sponsorship of a PatronRole, so that the risk-takers are guaranteed some minimum level of security even if they fail. However, they are not guaranteed the same level of rewards as people who succeed in lower-risk ventures; see CompensateSuccess.

✥ ✥ ✥

This pattern is a bit different from BuildPrototypes. Prototyping is one strategy towards running a SkunkWorks; however, a SkunkWorks project may just buy an existing product and integrate it with existing products or market it differently without doing any prototyping.

This pattern does not integrate with the other scheduling and organizational structures in the pattern language because it's a decoupled effort. The effort should evolve into a product over time and eventually incorporate patterns like SizeTheSchedule and SizeTheOrganization, but only after it's on its feet and has proven itself.

It is important that the SkunkWorks be organizationally separate from the mainline organization. This allows so-called disruptive technologies to flourish within the company. (For further information on disruptive technologies, see works by Clayton Christiansen, professor at Harvard [BibRef-Christianson1997].)

Though it's clear how SkunkWorks fit into a large organization, it can work on a smaller scale in small organizations as well. A couple of team members can develop innovative ideas "in the margins" as a side activity. This may be a particularly good outlet for employees whose skills are high enough that they seek challenges beyond those offered by day-to-day business.