Building IBM - The lesson of install base


An exceptional insight into the power of the install base.











Impact of Scale

  • Customer deliveries of the CPC began in late 1949, at which time more than twenty had been ordered by premier companies in airframe industry and by government agencies and laboratories. More than a year would pass before the first UNIVAC was accepted by the Census Bureau. Tom Watson was ecstatic. Nearly 700 CPC systems were delivered during the first half of the 1950's compared to only 14 UNIVACs during the same period. 

  • There is no simple way to compare the computational capabilities of these very different machines. The UNIVAC could execute a typical mix of arithmetic functions about twice as rapidly as the CPC. The UNIVAC also had an internal memory capacity more than one hundred times greater than the CPC, and it was equipped with magnetic tape storage devices that made it possible to move information in and out of memory rapidly. 

  • In the long run relative computational capabilities of these machines were less important than the number installed. During the first half of the 1950's, the installed UNIVACs satisfied computational needs at fourteen customer locations whereas the installed CPC's satisfied the needs at fifty times that many. The company was rapidly creating an infrastructure of knowledgeable customers, salesmen and servicemen for electronic computers. Contrary to widely held perceptions, IBM had the fastest start in the market for electronic computing capability. It did not have to catch up as most accounts of this era suggest. 

The UNIVAC was a far better machine with better functionality and internal capacity but CPC won the market because it had more machines in circulation (50x greater). 

Appealing to the mass market. 

  • By the time the last IBM 650 came of the manufacturing line in 1962, almost 2 thousand had been produced and installed. No other electronic computer had been produced in such quantity. 

  • Tom Watson Snr: "We had developed a machine for ordinary business that became the most popular computer of the 1950's"

IBM neglecting Software was their ultimate long term demise

  • Important to the success of RAMAC (Random Access Memory Accounting) was a novel method for finding stored information when its physical location on the disks was unknown. This first known use of randomization and chaining for storing and retrieving records was devised by H. Peter Luhn, who was already an established inventor when hired by Watoson Snr, in 1941. 

  • Luhns application for a patent on the method was rejected by the company's patent review board in Poughkeepsie in 1956 because the concepts were not embodied in hardware. The company had made a policy decision that computer programs and procedures were not patentable. 

Further Hardware inventions compounding IBMs competitive Advantage

  • It is difficult to overstate the impact the 350's disk technology has had upon the world in the years since its announcement... it is sufficient to say that flexible and fixed disk drives alone generated an estimated $12.5B in sales worldwide in 1983 - This invention was awarded an International Historic Mechanical Engineering Landmark by the American Society of Mechanical Engineers. 

IBM collected patents and continually invested in R&D for this reason. 

Industry Leading customer service helped win the market

  • Renting equipment and providing customer service at no extra charge became a cornerstone of watsons policies after he took over the company in 1914. Many a story is told of IBM service men working around the clock to get equipment running. If a fire, flood or disaster occurred IBM was ready to help. 

Patent monopoly was the source of market dominance and ultimately anti-trust behavior

  • Patents were of course the basis of IBMs early monopoly of the information processing business.

  • GE entered a Consent Decree in 1912 which essentially confirmed the legality of competitive advantages based on patents - this kick started the obsessive accumulation of patents. 

  • With the help of James Bryce, he built a strong patent portfolio. He identified and purchased relevant patents. He hired inventors and engineers to create new products and to file for patents. During the brief recession of 1922 Watson took advantage of an opportunity to buy the patents and research facilities and to hire key personnel of his most innovative competitor. 

The Anti Trust Act - Sherman - A really important case study for current day monopolies 

  • Watson Dominated the information processing industry so much that in 1952 the US department of justice filed an anti trust suit alleging that IBM no owns 90% of all tabulating machines in use in the USA. 

  • A reading  of all past court decisions regarding the Sherman Act proved that a defense was unlikely to succeed unless the company could prove 1) Company does not possess Monopoly power or 2) Monopoly achieved unintentionally  

  • The First anti trust suit was initiated in 1952 and settled by a consent decree in 1956 (4 years!) - IBM had to allow access to its patents to anyone that requested them. If true competition didnt emerge in 7 years then IBM needed to divest itself back down to 50% market share. 

Moving onward to next generation technology 

  • in 1965 the production of SLT modules (Semiconductors) reached 36M, and 90M in 1966 this was more semiconductor devices than were produced by all other companies in the world combined! 

  • Because people were rushing to produce volume of semiconductors they forgot to patent the processing machinery/equipment. This is what commodities Semiconductors today and the reason why they are produced so cheaply everywhere in the world. 

Moores Law creating a headwind

  • in 1970 the cost of hardware was dropping that rapidly that people thought that new customers and applications would have to be found at an alarming rate just to maintain the same dollar value of sales each year

Breaking up the company is better than a consent decree

  • Considering all the heavy constraints anti trust enforcement placed on IBM's ability to compete, it might have been wiser to break the company into several smaller units as the government proposed. Support for this view is provided by the experience of AT&T. Yielding to government pressure it agreed to divest itself of its local telephone business in the same year that IBM "won" its antitrust case. After that the value of AT&T stock more than doubled whereas the value of IBM stock declined. "I am convinced", the CEO of AT&T told shareholders in 1994, "that the breakup forced us to make changes in the 1980's  that prepared us for the competitive challenges we are confronting in the 1990's"/ 

Breaking up a monopoly seems to deliver more value - Standard Oil and AT&T.