FOOL ON THE HILL: An Investment Opinion
How Xerox Forfeited the PC War

The fundamental tension between innovative commando employees and bureaucrats can be managed by using "think tanks" to generate innovative ideas. Historically, companies such as Xerox, IBM, and AT&T have had mixed results using this technique.

By Rob Landley (TMF Oak)
September 18, 2000

This is Part 4 in an ongoing series. Part 3, How Companies Grow Up, ran last week. Part 2, Berkshire's Sustainability, ran earlier this month. Part 1, How a Start-up Evolves, ran in July.

This column revisits author Robert X. Cringely's commando/infantry/police metaphor covering the types of employees companies depend on as they mature. Commandos are the backbone of any start-up, swashbuckling innovators who are unpredictable and uncontrollable yet unbelievably productive. Infantry are the obedient workers who methodically grow a company from its IPO to market dominance. Finally, police are the bureaucrats and middle managers who defend the entrenched position of an established market leader.

The fundamental tension here is that commandos can't stand police, and vice versa. Sticking a mad scientist in a department meeting isn't going to work. Creative individuals cannot accept rote procedure, and bureaucrats cannot accept loose cannons who don't conform to "The [company name] Way." When "the suits" take over, the commandos leave. Every time. Yet without innovative dreamers, innovation is virtually impossible.

The only really effective way to innovate is to use commandos, and the only way to keep commandos around is to isolate them completely from the stifling influence of bureaucracy. Take a physically isolated location and staff it with infantry -- who are tolerable, if not exactly liked by commandos and police -- and have them hire commandos who are protected from the rest of the company's bureaucracy.

This is where AT&T's (NYSE: T) Bell Labs, now Lucent Technologies (NYSE: LU), comes in. Other examples are Xerox's (NYSE: XRX) Palo Alto Research Center (PARC) and various scattered IBM (NYSE: IBM) facilities such as the one in Boca Raton, Fla., which produced the IBM Personal Computer. These organizations are commonly called "think tanks," as they isolate commandos in a suit-proof environment.

The job of a think tank is to perform "basic research," which basically means anything the commandos feel like doing no matter how useless it seems at the time. Give the commandos room to play and see what they come up with. This can lead to amazing, though often annoying, results.

Annoying because the "let them innovate and figure out what to do with it later" approach works pretty well up until the word "and." The problem is that the level of insulation required to prevent the suits running a mature third-wave company from cramping the commandos' style also tends to prevent the third-wave company from harvesting the fruits of their labors. (This is partly a matter of mindset: Even when bureaucrats have the future in their hands, they just don't see it, as has been proven time and again.)

AT&T's fish bowl, full of carefully insulated commandos in Murray Hill, N.J., produced what is arguably the greatest invention of the last century: the transistor. AT&T was happy to upgrade its telephone switching network with transistors and leave it at that. Dozens of start-up companies saw far more potential in the transistor, licensing it from AT&T and forming companies with names such as Texas Instruments (NYSE: TXN), Sony (NYSE: SNE), Hewlett-Packard (NYSE: HWP), and Intel (Nasdaq: INTC).

A couple of decades later, the same Bell Labs produced the Unix operating system more or less by accident, which AT&T used in its phone switches and a few in-house applications but otherwise ignored. Once again, the rest of the world took the new technology and ran with it. Any modern operating system that uses "subdirectories" (sometimes called "folders") probably copied the concept from Unix -- and yes, this includes both Microsoft's (Nasdaq: MSFT) DOS/Windows and Apple's (Nasdaq: AAPL) Macintosh.

By 1984 Unix had changed the entire software industry so profoundly that AT&T consented to a government breakup in large part to get permission to enter the computer business and profit from Unix commercially. Losing its long-distance business was a small price to pay to gain ownership of the computer industry through Unix. (AT&T totally screwed up this opportunity, but that's another story.)

Similarly, Xerox PARC invented modern desktop computing. Windows, icons, mice, pulldown menus, "What You See Is What You Get" (WYSIWYG) printing, networked workstations, object-oriented programming -- the works. Xerox the copier company feared the paperless office and formed a think tank to invent it before anybody else could, but once its commandos had succeeded, it simply couldn't bring itself to disrupt its core business of making copiers.

Xerox could have owned the PC revolution, but instead it sat on the technology for years. Then, in exchange for the opportunity to invest in a hot new pre-IPO start-up called "Apple," the Xerox PARC commandos were forced -- under protest -- to give Apple's engineers a tour and a demonstration of their work. The result was the Apple Macintosh, which Microsoft later copied to create Windows.

On the other hand, Xerox did put a windowing interface on a little display on the side of some of its copiers.

Apple's IPO scared the heck out of IBM, which wanted to flood the new "Personal Computer" market niche with a product of its own to squash the competitive threat of the Apple II. Unfortunately, an internal audit of IBM's bureaucratic processes had just revealed it took IBM nine months to ship an empty box. If the product wasn't ready in a year it would be too late to threaten Apple. Completing the new project in time, while imperative, seemed impossible.

IBM's Boca Raton team volunteered, with the condition that management leave them alone for a while and look the other way while they violated every procedure in the book. The result was the original IBM PC, which again spawned an industry of clones.

IBM think tanks have also produced the IBM Microdrive, the scanning tunneling microscope, and countless technical innovations in data storage and circuitry design and manufacturing. Relatively speaking, IBM actually has a pretty good track record of commercializing the work of its pockets of commandos, which means it is only hideously inefficient about it rather than completely wasteful. "Good" in this context is still pretty bad.

Through think tanks, third-wave companies can sponsor real innovation. Unfortunately, controlling the direction of that innovation is almost impossible, and knowing what to do with that innovation once it's been produced is equally difficult for the bureaucratic mind. Think tanks are good for the industry as a whole, but often benefit other companies as much as the think tank's sponsors. For many companies that admit their inability to exploit the opportunity of having first crack at unproven new technologies, donations to universities serve much the same purpose with less hassle.

-- Oak