The Giant's Small Computer
You are an executive at the world's dominant computer company. Small personal computers made by start-ups are selling fast, and your company has nothing to offer. You have been asked to get a product to market in about a year.
How do you build a personal computer quickly enough to matter, without giving away the future?
The situation
It is 1980. Your company dominates computing. Its large systems run the world’s banks, airlines and governments. But a new market has appeared that it does not serve: small personal computers, sold by young companies to hobbyists, small businesses and schools. They are selling in growing numbers, and some business customers are beginning to buy them.
Your company’s normal development process is thorough, slow and built around proprietary technology: its own chips, software and parts, designed in-house and tested exhaustively. A product of this kind would normally take several years. Senior management wants a personal computer on the market in about a year, and has authorised a small team, located away from headquarters, to find a way.
What you know
Personal computers from start-ups are built largely from components anyone can buy: microprocessors from chip makers, memory, disk drives and keyboards from suppliers. Their success depends heavily on software and add-on cards made by other companies, which in turn depends on those companies knowing how the machine works.
A capable 16-bit microprocessor is available from an established chip maker. A small software company has offered to supply an operating system, which it would obtain and adapt; it wants to keep the right to license the software to other manufacturers too. Your brand would reassure business buyers, and your sales force reaches every large company.
What you don’t know
You do not know how large this market will become, whether it will matter to your core business or whether your company’s culture can produce a small, cheap product at all. You do not know how other manufacturers might respond if your design became popular, or whether the small software company and the chip maker will remain small.
The voices in the room
Traditionalists argue that the company’s strength is its own technology, and that a machine built from other people’s parts would be unworthy of its name and impossible to defend. The small team argues that speed matters more than control, that an open design will attract the software and add-ons a personal computer needs, and that the company’s brand and distribution will carry the day. Others think the whole market is a distraction.
What happened
The company was IBM. Its small team in Boca Raton, Florida, built the IBM Personal Computer from outside parts: an Intel processor, an operating system licensed from Microsoft, and components from many suppliers. IBM published the technical details to encourage add-on makers. Microsoft’s licence was non-exclusive, so it could sell its version of the operating system, MS-DOS, to others. The PC was announced in August 1981, with a base price of $1,565, and sold far beyond expectations. It became the standard for business computing.
The same features that made it a success let others share in it. Compaq and others built compatible machines; the one proprietary component, the BIOS, was legally replicated through clean-room engineering. Intel and Microsoft supplied the whole industry and became enormously valuable. In 1987 IBM tried to regain control with the PS/2 line and a proprietary expansion bus, and, with Microsoft, the OS/2 operating system, with limited success. In 2005 IBM sold its PC business to Lenovo.
Judging the decision apart from the outcome
It is easy, in hindsight, to call this the most expensive mistake in business history: IBM gave away the processor and the operating system, the two parts that turned out to matter most. But the symmetrical question is what else could have worked in 1980. A proprietary in-house machine would probably have arrived years later, into a market others had shaped, without the software and add-ons an open design attracted. The open approach is why the PC became the standard at all.
The weakness in the process lay in the terms rather than the strategy. Accepting a non-exclusive operating system licence and relying on a single outside processor supplier left IBM without control of the parts that every compatible machine needed. In transaction cost terms, the processor and operating system were the specific, scarce assets of the new standard, and they belonged to others. Some of this was knowable: the software firm’s wish to license to others was explicit, and the logic of compatible machines was understood. What was not knowable was how large the market would become, which is what made those terms so costly.
What a good process would have weighed
A good process would have kept the open design, which won the market, while seeking more control over the scarce parts: an exclusive or longer-term licence, ownership of the operating system or a stake in the suppliers. Whether such terms were obtainable at the time, at a price IBM would accept, is uncertain. That is why the grade is Sound rather than Strong: the strategy was right, and the terms underpriced a future that was hard to see.
Sources and limitations
The launch date and price are from IBM’s own history pages. The account of the team and the licensing terms draws on published histories of IBM and the personal computer industry; participants’ accounts differ in details, including how the operating system deal was reached, which is not described here. The judgement about what better terms were obtainable is the author’s and is uncertain.
The reveal
IBM, 1980
Assemble it from outside suppliers' parts, publish the technical design, and license an operating system from a small software firm.
Mixed
The machine launched in 1981 and set the standard for the industry, but its open design and non-exclusive licences let clone makers and the suppliers of its processor and operating system capture most of the long-term value. The company sold its PC business in 2005.
Sound
Judged on what could be known at the time, separately from the outcome. See moral luck.
The forces pulling at the people deciding
-
Buying parts from the market was far faster than building them, and the cost of dependence on suppliers seemed small when the market was small. The supplier relationships later turned out to be where the value concentrated.
-
Irreversibility hidden in openness
Publishing the design and accepting a non-exclusive software licence were hard to reverse. Once the standard spread, attempts to take back control met resistance from the very ecosystem that openness had created.
-
An organisation built for something else
A small, separate team was set up away from headquarters precisely so it could act unlike the rest of the company. Its product reflected that independence, for good and ill.
-
Once software and add-ons were written for the design, the standard locked in, independent of who made the machine.
Signals that were visible at the time
- The software firm supplying the operating system kept the right to license it to other manufacturers.
- The processor came from an outside chip maker who could sell it to anyone.
- The design was published to attract add-on makers, so others could build compatible machines if they could replicate the remaining proprietary parts legally.
- Start-ups were already selling personal computers successfully with off-the-shelf parts.
Openness can win a market; whoever controls the scarce, specific parts of an open standard keeps the value.
Follow the thread
Connections
Where this comes from
Sources
- IBM, “The IBM PC's debut,” IBM History.
- Paul Carroll, Big Blues. The Unmaking of IBM (Crown, 1993).
- Paul Freiberger and Michael Swaine, Fire in the Valley, 2nd ed. (McGraw-Hill, 2000).
- James Chposky and Ted Leonsis, Blue Magic. The People, Power and Politics Behind the IBM Personal Computer (Facts on File, 1988).
Teaching this case? The teaching note has learning objectives, a 60-minute session plan, discussion questions and the key facts on one page.
Shrikant Kumar. “The Giant's Small Computer.” Decide case, Telos. https://shrikantk.com/decide/the-giants-small-computer/