MANAGING INNOVATION: A Company Organized to Reflect the Operation of its Products
Ken Olsen was excited by the opportunity to work with one of the first mainframe computers. As an inquisitive engineer he was frustrated, however, by the hands-off environment that IBM maintained with its Whirlwind computer. In many different ways both the IBM organization and the operation of their equipment reflected their distinctive top-down organizational architecture and the permission-based use that Olsen did not like. He felt it was an unnecessary constraint on learning what computers could do, and upon their further development.
Olsen deliberately designed a more open and egalitarian company. Digital Equipment Corporation was remarkably flat as an organization. Each new idea was presented to an Operations Committee to be critiqued and defended. Cross-department teams formed around interesting projects. Olsen believed this peer review would lead to the best, most elegant solution.
|Rigid, centralized chain of command
||Decentralized peer review
||Interactive time-sharing minicomputers
|Networking by permission
||Networking as equals
And there were many innovations, especially in networking mini-mainframe computers:
- First computer company to connect to ARPAnet
(previously populated by government and research agencies)
- First computer company with a registered domain name. Fifth in line all together.
- DEC had the first corporate gateway (e-mail over the Internet)
- Did champion Ethernet with Xerox and Intel, but gets the real credit for making it financially viable.
- First live streaming over Web (Mars Pathfinder mission) reached 1,000,000+ users.
- First to bring real-time election results to the Internet
- Created the first Internet firewall
- First comprehensive Internet search engine (AltaVista)
- First widely available text-to-speech system (Stephen Hawking used a DECtalk until 2009)
- First to deliver remote software updates
An organization designed to encourage innovation creates a culture that values innovation. Once approved by the Operations Committee, the person proposing a winning idea could form their own team to develop a new product line.
This freedom to pursue a new approach is supported by sales of minicomputers that, while a tenth the cost of a mainframe, were still out of the reach of an individual user.
But a matrix-based organization can be difficult to manage. As the industry shifted to cheaper, less elegant home PCs and shareholders looked for their own dividends, the dollars left to support product development became limited. Projects and their teams began to compete within DEC for scarce dollars. Decentralized decision-making can lead to costly duplication and fighting for turf. Once Olsen left DEC in 1992, its successful product lines were sold separately for cash and what remained was merged with Compaq.