A Computer Called LEO: Lyons Teashop and the World's First Office Computer (review)

In what was possibly the ultimate in customer-driven innovation, J. Lyons & Company developed, built, and operated the world's first business computer, LEO (Lyons Electronic Office), after World War II. Lyons took this unusual path because of its immense calculating needs, a consequence of its successful growth into the largest British food manufacturer, distributor, and retailer; a lack of machines (and interest) from possible British sources; and a long-engrained drive for greater efficiency by a progressive company and one manager in particular, John Simmons.

Simmons, a Cambridge wrangler, joined Lyons in 1923 and introduced scientific management, albeit with concern for employees, and other innovations, including the world's first commercial use of microfilm in 1935. After World War II, he convinced the firm to build a computer to handle its complex clerical and accounting needs.

Lyons funded and participated in Maurice Wilkes' construction of EDSAC, a path-breaking stored-program computer, from 1947 to 1949 at Cambridge University. This success gave Simmons the confidence to propose that Lyons build its own computer in 1949, based on EDSAC. On November 29, 1951, LEO ran a routine office task, the weekly bakery valuation, in five hours, compared to fifty hours by clerks. At that time, LEO was one of only three computers in Britain, according to Georgina Ferry (the other two are not specified, so military computing might be excluded).

Two major challenges were developing reliable peripherals and ensuring the reliability of components, especially valves (vacuum tubes), and the entire system. The peripheral problem was mainly a failure of British firms to produce as promised, leaving the LEO team scrambling to compensate. In December 1953 LEO was officially completed and in February 1954 Lyons introduced it to the world. The firm gradually increased its operations, moving from payroll to ordering food for the teashops. Significantly, the programmers did not simply automate existing procedures but analyzed the flows of information (including asking everybody involved what would help them) and restructured them. Efficiency came through smarter as well as faster processing.

LEO was a success because its designers focused on what the customer—in this case, the people building it—actually wanted and a determination that the best was the enemy of the good. LEO was not the most efficient computer, but it was a technology designed for businesses.

Even before LEO was complete, Simmons proposed building and selling better LEOs. Lyon's board, facing postwar financial challenges, hesitated at the large cost before setting up a subsidiary in November 1954. The new Leo Computers proved less adept at building new generations of computers and marketing them. LEO II was a major advance—but it did not start working until May 1957, not the planned December 1955. The company lacked resources, and its handicrafts approach to construction, while normal for a research project, was not suitable for large-scale production. The major lack of resources was qualified people—LEO II's engineering team comprised only a score of people. Obtaining adequate peripherals, especially a high-speed alphanumeric printer, remained a problem. Only eleven LEO IIs were built. LEO III, which used its own programming language instead of the common COBOL, was also technically impressive and lacked resources. Like LEO II it entered a growing market in 1962 but one with increasing numbers of machines from other competitors, many better known and with more resources. LEO's narrow window of opportunity had closed. Only eighteen LEO IIIs were sold.

Within a decade, LEOs and British office computing had vanished under the weight of offerings from IBM. Georgina Ferry emphasizes the importance of the American government's promotion of computing by encouraging open communication, as well as by offering $400 million in contracts to IBM in the 1950s alone. In contrast, the British government was obsessive about security, focused money on...