The First Business Computer: a Case Study in User-Driven Innovation

Frank Land The first business computer: a case study in user-driven innovation

Article (Accepted version) (Refereed)

Original citation: Land, Frank (2000) The first business computer: a case study in user-driven innovation. IEEE annals of the history of computing, 22 (3). pp. 16-26. ISSN 1058-6180 DOI: 10.1109/85.859523

This version available at: http://eprints.lse.ac.uk/19150/ Available in LSE Research Online: September 2012

LSE has developed LSE Research Online so that users may access research output of the School. Copyright © and Moral Rights for the papers on this site are retained by the individual authors and/or other copyright owners. Users may download and/or print one copy of any article(s) in LSE Research Online to facilitate their private study or for non-commercial research. You may not engage in further distribution of the material or use it for any profit-making activities or any commercial gain. You may freely distribute the URL (http://eprints.lse.ac.uk) of the LSE Research Online website.

This document is the author’s final manuscript accepted version of the journal article, incorporating any revisions agreed during the peer review process. Some differences between this version and the published version may remain. You are advised to consult the publisher’s version if you wish to cite from it.

The First Business Computer: A Case Study in User-Driven Innovation

Frank Land London School of Economics Nearly half a century ago the world’s ﬁrst business computer application was rolled out. The host for the application was a British catering and food-manufacturing company, which had developed and built its own computer, designed for business data processing. The author traces the endeavor’s history and presents an analysis of how and why the company—J. Lyons—was in a natural position to take on the challenge, the precursor of the information revolution we see today.

The final years of the 20th century mark the company achieved precedence in a field where fifty-year history of the “invention” of elec- it might be expected that high-technology tronic computing. Although the United States equipment manufacturers would be expected is now the greatest and most influential user of to lead. In the second stage, some theoretical computer technology, and has and continues frameworks derived from recent research into to produce great innovators, many advances company culture and corporate success are had their source in Europe. Today Europeans examined to see if they throw further light on continue to play an important role in the evo- the case. lution of computing. One of the most interesting innovations to The case study come first from Europe was the application of This study takes the case of J. Lyons to the computers to automating business and admin- point where the company decided to design istrative procedures and solving business prob- and build their own computer in 1949. (Much lems. Today, business data processing (or of the material in this case study, including management information services as it’s often most of the quotations, are taken from Bird,3 called in the US) in all its forms, ranging from and Caminer et al.4) transaction processing and decision support to desktop publishing and e-mail, has become the Company background largest single use for computing technology. J. Lyons was founded in the late 19th cen- The use of computers is pervasive in all kinds tury by the successful tobacconists Salmon and of businesses and administrations. Gluckstein. The company name is derived from One of the first uses of computers in the a relative of the founders because, at the time, business area came from a British food manu- Salmon and Gluckstein were not confident that facturing and catering company, J. Lyons and they would make a success of the new enter- Company, and dates back to 1949.1 The case is prise and did not want their own names asso- of particular interest because this innovative ciated with a risky venture. use of computers was driven by an organization Salmon and Gluckstein formed the compa- that had no prior experience in electronics, yet ny in 1887 to fill a gap in the catering market. they built and applied to their business one of They had noted that catering facilities at exhi- the earliest stored-program computers, the LEO bitions and trade shows were often of a very I. Truly a case of user-driven innovation. poor standard. They thought there would be a This article sets out, by means of a case market for affordable, quality food combined study and an analysis of the case, to describe with some entertainment at the many shows how J. Lyons came to design and build the and exhibitions held in locations throughout world’s first business computer.1-5 The analysis the country. The provision of catering facilities is carried out in two stages. In the first, features for short duration exhibitions required metic- of the case are selected that might provide a ulous planning both in the sense of controlling pragmatic explanation of how a food business the operation with its highly-peaked activities

16 IEEE Annals of the History of Computing 1058-6180/00/$10.00 © 2000 IEEE so as to provide the quality demanded by the ees such as George Booth, the company secre- public and at the same time to maintain tight tary—recognized that progress could only be control of costs. maintained if the company hired people of Thus J. Lyons management from their earli- high academic standing. In 1923 they hired est days developed a culture of seeking the John Simmons direct from Cambridge univer- most effective ways of controlling their busi- sity (where he had been a Wrangler, the term ness. The company expanded rapidly from suc- Cambridge used for the highest performers in cess at catering for special events, very much a the mathematics degree, and where he had niche market, to becoming one of the most earned a first class honors degree in mathe- important and respected food companies in the matics) to join the company as a management UK. In the 1940s and 1950s J. Lyons was one of trainee and statistician. the most successful businesses in the UK with In an interview1 with Chris Evans at the its products, services, and establishments— Science Museum about 1970 Simmons Lyons Tea, Lyons Cakes, Lyons Ice Cream, the explained: catering services provided at great national events such as the Buckingham Palace Royal In fact I was engaged to try to build up a system Garden parties and the Wimbledon tennis of information for the management of the com- championships, and the Teashops and Corner pany which would be superior, more sensible, Houses (large, popular, but stylish London than just depending upon the profit and loss restaurants)—being household names. account and such like. … in this respect the com- From its beginnings the company embraced pany was already ahead. a culture of self-sufficiency. It built or acquired through purchase the capability to provide the Simmons rose rapidly in the management goods and services needed for the core business hierarchy and was responsible for introducing a of catering. It became a vertically integrated number of far reaching changes in practice and company, with its own engineering works in organization. These changes were designed designing and building machinery needed by to cope with the growth of the business and to the factories, body shops that built the compa- achieve the control necessary for the mainte- ny’s transport fleet, laboratories engaged in nance of profits. Simmons ensured that the food research, and laundries washing the com- company kept up a continuous striving for pany’s dirty linen. Of course, where the com- improving practice and, in doing this, intro- pany identified suppliers capable of meeting duced many innovations in particular to the Lyons’ exacting standards they were used, but management support services. This enabled the company ensured that such suppliers main- management to keep their business under con- tained laid down standards. stant cost effective review. Nevertheless, the culture of self-sufficiency In the early 1930s Simmons set up the led to a belief by management that there was Systems Research Office whose function was to little the company could not do as well or bet- review existing systems; trawl the world for bet- ter than any outsider. Wartime experience ter way of supporting management activities; enhanced that confidence. Some of the Lyons and invent, test, and implement improve- factories had been converted to the making of ments. The Systems Research Office, working munitions and company management prided with line managers, produced a stream of busi- itself on their efficient conversion that put the ness process innovations from the time of its company on a wartime footing. establishment. Examples include the notion Lyons had built its success on quality prod- that sales representatives, each having a cus- ucts and services sold to a mass market. Selling tomer group of many small retailers, would not to a competitive mass market required tight only be responsible for selling to their own cus- control over costs and margins and a sensitive tomer group but would also be totally respon- response to customer preferences and market sible for the accounting, credit, and payment movements. Given the very large number of functions. These functions were conventional- daily transactions—counted in tens of ly carried out at arms length by a separate thousands—with an average very low value and accounting office. The introduction of “travel- a profit margin per transaction of a fraction of er covered credit” was a radical business process a penny, the company could progress only by a innovation, which increased efficiency and the constant striving for value adding innovation. effectiveness of the representative—the sales Senior management—primarily members of representative took responsibility for the the founding families, the Salmons and accounting necessary to maintain a customer Glucksteins, plus a few trusted senior employ- account.

July–September 2000 17 First Business Computer

Before World War II the Systems Research undertake investigations of the “what if” Office had for many years been investigating type—suppose we wish to increase the produc- the possibility of coping with the mass of trans- tion of Swiss rolls by 10 percent and reduce the action data by some kind of mechanization or production of cupcakes by 3 percent, what automation. They had started to investigate the would be the effect on gross profit? The possibility of devising a document reader for arrangement ensured direct access by senior transaction data. They had researched the pos- management to information originating at the sible application of a unit record systems based operating level and bypassed the more usual fil- on punched cards but rejected these as too tering through layers of middle management. localized, too constraining, and too costly. This management style provided the com- Lyons had only one punched card installation pany, long before the advent of computers, with and that had a very limited application. both an almost real-time management infor- Instead, the company had installed alternative mation system and a decision support system of types of office mechanization based on considerable sophistication. In addition, the accounting machines and calculators. architecture and system provided senior man- By 1946 Simmons had been promoted to agement with a detailed picture of the week’s the post of chief comptroller responsible for all trading on the Monday of the following week. the management accounting and support func- Yet in many other ways the company was tions in the business, and in 1950 he joined the deeply traditional and conservative. It operat- family dominated Lyons’ board as an employee ed on a strictly hierarchical basis. At the top director. In 1932 Lyons engaged another were the owners, the founding families. They Cambridge Wrangler, T.R. Thompson, who had ran the company with the help of very few previously been employed as assistant secretary employee directors. Each grade of management to a large Liverpool store. Under Simmons’ had its own dining room. Separate toilets divid- tutelage Thompson also rose rapidly in the ed managers from the rest. As might be expect- organization and by 1946 had reached the posi- ed in a company with these types of tion of chief assistant comptroller. relationships, interaction between employees John Kay6 in his study of what makes busi- was conducted in a formal way with formal nesses successful suggests that architecture, by modes of address, particularly to one’s superi- which he means the structure and style of ors in the hierarchy. Sports were encouraged responsibilities and relationships, is one of the and a great variety of team games took place important enablers. Over the years Lyons had with enthusiasm on the company’s large sports developed a distinct management style based grounds. Trade unions were discouraged, on the way information was passed from oper- though the family took a paternalistic interest ations—manufacturing, selling, distribution, as in its staff. well as the concomitant operations concerned Nevertheless, top management encouraged with invoicing and payments—to the decision innovation in working practices7 whether in making senior management. Each of the many the factory with the design and introduction of businesses (tea, teashops, ice cream, bakeries, continuous ovens for the manufacture of miles kitchens, outside catering, and so on) had their of Swiss rolls per day, or in the office with the own groups of clerks and managers. The vast adoption of microfilm to provide a single mass of transaction data stemming from these record of orders received that served as delivery operations was summarized and compared note and invoice. Management insisted, how- with preset standards, forecasts, and budgets. ever, that all new practices had to be tested and The resulting information was analyzed by the prototyped until most elements of risk had junior manager in charge of each group, who been eliminated. would be responsible for explaining any important variances. (In the Lyons grading structure, Initiation junior managers were placed in the F or super- In May 1947 two senior company execu- visor grade and not deemed managers. Many of tives, Thompson and Oliver Standingford— them rose to manager rank in their subsequent another assistant comptroller who had joined careers and some reached the very top of the Lyons in 1930 as a management trainee—visit- organization.) The junior manager had a direct ed the US to see what developments in office line to the senior manager, often a Lyons direc- equipment had taken place since the start of tor, responsible for that activity and to whom World War II. They found little significant the functioning of that activity had to be change in office equipment. But they did hear explained. At the same time the senior manag- about experiments with electronic brains, such er could ask the junior liaison manager to as the engineering numerical integrator and

18 IEEE Annals of the History of Computing computer, or ENIAC, used exclusively for mili- tary, scientific, and engineering calculations. They recognized quickly that the characteris- In May 1949 EDSAC would tics of the machines they saw could be modi- fied to provide the necessary capabilities to become the world’s first solve some of the problems of business data processing. They visited a number of the pio- stored program computer neers and were particularly impressed by Howard Aiken at Harvard whose practical to do regular appreciation of bringing equipment into reli- able operation mirrored their own ideas of the mathematical work. problems that they might face Thompson and Standingford discovered that few of the pioneers had thought about the possibility of applying the machines to busi- that included inter alia the following statement: ness operations, though American Prudential Insurance Company had already established We believe that we have been able to get a plans to use a computer for premium billing glimpse of a development which will in a few and actuarial calculations. Both Aiken at years time, have a profound effect on the way in Harvard and Goldstine at Princeton told them which clerical work (at least) is performed. Here, that England’s Cambridge university under for the first time, there is a possibility of a Professor Douglas Hartree—a mathematical machine which will be able to cope, at almost physicist—had started work on the electronic incredible speed, with any variation of clerical delay storage automatic calculator (EDSAC). In procedures, provided the conditions which gov- May 1949 EDSAC would become the world’s ern the variations is predetermined. What effect first stored program computer to do regular such machines could have on the semi-repetitive mathematical work. work of the office needs only the slightest effort On their return to England Thompson and of imagination. The possible savings from such Standingford visited Cambridge where they a machine should, be at least £50,000 a year. The were introduced to Maurice Wilkes, director of capital cost would be of the order of £100,000. the mathematical laboratory, who had started We feel, therefore, that the Company might work on the EDSAC. They were most impressed well wish to take a lead in the development of with the advances in the technology, well the machine and indeed that, unless organisa- beyond anything they had seen in the US. tions such as ours, the potential users, are pre- However they wrote: pared to do so, the time at which they become commercially available will be unnecessarily Progress on the machine is slow and they talk of postponed for many years. 12-18 months before completion. They have all their plans drawn and the hold-up is purely due The report describes details of Thompson and to lack of money. Dr Wilkes and one draughts- Standingford’s visits and the ideas discussed man assistant alone are handling the job, assist- with some of the computer pioneers. It describes ed at the moment by two vacation students. We how the computer works, suggests how the told them that unless they could proceed more mathematical computers might need to be mod- rapidly they may find the machines for sale in ified for business use, and goes on to indicate a America before they complete their first model. number of possible applications and how they We were told that given £2000 they could com- might be handled on the computer. These plete much more rapidly. Both Professor Hartree included ideas for a variety of transaction pro- and Dr Wilkes were willing and keen to co-oper- cessing applications, but also hint at the possi- ate with us, in particular they are interested in bility of using the computer for word processing. applying their machine to any clerical job we Under the heading “Steps that might be may suggest.8 taken by Lyons to advance the development of Electronic Machines” they wrote: Having gathered all the information they thought they needed, Thompson and Our first concern is, of course, the advantage Standingford produced a report for the Lyons which Lyons may gain from the commercial board, which they first showed to Simmons. development of electronic machines, but there With Simmons full backing they presented the is a wider aspect which cannot be overlooked. report to the board, with a cover memorandum This machine may well be a prime factor in

July–September 2000 19 First Business Computer

relieving the present economic distress of the build their own computer and to help recruit a country. In this respect we cannot help but feel chief engineer, Dr. John Pinkerton, to head the that Lyons occupies a key position; no-one else technical side of the project. A project team was here, as far as we can learn, has realised the far- established and work commenced in 1949 to reaching possibilities of electronic machines. design and build the computer. Simmons sug- We assume that Lyons will want to take full gested a name for the machine—Lyons advantage of these machines for their own Electronic Office, or LEO. offices. It is possible for us to play a passive role LEO differed from EDSAC in that it was by merely keeping in touch with developments, designed specifically for large-scale data pro- and in due course buying machines as they cessing with multiple buffered input and out- become available, probably from American put channels, capable of being linked to a sources. But such a role would not enable us to variety of input and output devices. By 1951 have any influence on the kind of machine built, the LEO team had built the basic computer and and without commercial influence they may be the first experimental, but regular, business built in a form more suited to handling mathe- application was being run for the Lyons bak- matical and census calculations owing to the eries. Further work in linking input and output influence of the large governmental concerns. devices, extending the store, and improving reliability continued until in December 1953, Finally, the report laid out five possible LEO I was formally declared completed and strategies for the Lyons board to consider that ready for full-scale operational work. would enable the company to apply computer technology. These included the suggestion that Discussion and analysis Lyons should support the efforts of Wilkes at To many observers the notion of a catering Cambridge with a financial grant and at the company in England shortly after World War II same time provide him with examples of cleri- successfully pioneering what is now called the cal procedures to use in programming experi- information revolution seems unlikely and ments. An alternative strategy suggested that: incredible. Lyons not only recognized the potential role of electronic computers for busi- We could build a machine in our own workshops ness data processing but actually designed and drawing on information and advice from built such a machine. Indeed, the US edition of Harvard University. the book User-Driven Innovation: The World’s First Business Computer4 is called LEO—The The board was persuaded to accept the basic Incredible Story of the World’s First Business recommendation that Lyons should itself take Computer.1 the initiative to acquire a computer for its office work. Once Booth, the company secretary, had The pragmatic explanation thrown his support behind the recommenda- Yet looking at J. Lyons in the period after tion there was no dissension among board World War II, their recognition of computers as members. They agreed to support the a way ahead and the decision to build their Cambridge venture but to work towards build- own should not be regarded as surprising. ing their own machine with the help of Wilkes There are a number of features in the way the and his team. The board also agreed to help company had developed and was organized in finance the work at Cambridge with a grant of the immediate postwar period (1945-1955) that £3000, the money to be made available as soon provided the basis for the pioneering move as Cambridge could demonstrate that the into business computing. EDSAC could carry out serious mathematical work. In addition, Lyons offered to make one The nature of the business. The business of their own people, Ernest Lenaerts, who had was characterized by the very large range of been a clerk in the Lyons offices prewar but had food products and food providing services. This worked on radar during the war, available to was supported by a range of company owned Cambridge. Lenaerts, whose formal education and managed services. The company sold its has finished with high school examinations products directly to the retail trade or through (matriculations), became an important con- its catering outlets directly to the public. This tributor to the development of EDSAC and later resulted in a mass of transactions with relative- a key person in the design and building of ly low average values. Competitiveness depend- Lyons’ computer. ed on advantageous pricing of its products. In return for the Lyons help package, Thus the company was constantly fighting to Cambridge agreed to help Lyons design and contain or reduce its costs of doing business—

20 IEEE Annals of the History of Computing its transaction costs in the parlance of econo- mists. Hence, its constant search for improving its business processes in all aspects of the busi- By recruiting people with top ness including manufacture; distribution; supply; and most relevant to the decision to academic qualifications and experiment with electronic computers, the support services provided by the office. The latter permitting them to take were responsible for the standard processes associated with bookkeeping and accounts, leadership roles, the Lyons payment of suppliers, charging customers, and paying the staff but also with the management management ensured that it support activities of providing management information. By the middle 1930s the compa- was never short of ideas on ny had established a systems research office to put the search for improved methods onto a how to improve. more formal footing.

Personnel and leadership. Senior management recognized that to sustain the growth and competitiveness of the company it had to have trol by becoming as self-sufficient as possible. top-class management. Lyons was one of the The second strategy, which he calls markets, earliest companies to establish a management involves going to the market (or outsourcing in trainee scheme and many of the people later today’s parlance) for all but the core activities involved with the LEO enterprise first joined of the business. Companies minimize costs by Lyons as management trainees. These included using the mechanisms of the marketplace and Standingford and Thompson. David Caminer, retain control through contractual arrange- probably the world’s first business program- ments with suppliers. ming manager, joined Lyons as a management Lyons, in general, preferred the first strategy trainee in 1936, became head of the systems because they had the confidence and experi- research office, and joined the LEO team as one ence to believe they could provide the relevant of its first members. goods or services more effectively than any By recruiting people with top-class academic contractor. Of course, some goods and services qualifications, such as Simmons and Thompson, had to be subcontracted or purchased when and permitting them to take leadership roles in investigations showed that appropriate suppli- the management of the company, the Lyons ers were available. But even then, Lyons, like management ensured that it was never short of Marks and Spencer (one of the UK’s leading ideas on how to improve. This was reflected in retailers), set up an organization that set stan- the esteem in which the company was held in dards and ensured that the suppliers them- the first half of the 20th century. Their leading selves were efficient and were always seeking role in developing systems was widely acknowl- improvements. Thus it did not take Lyons long edged beyond the boundaries of their industry to realize that outside suppliers capable of sector. Simmons played a leading role in the meeting the requirements envisioned by establishment of the Institute of Administrative Simmons and Thompson simply did not exist Management and was elected as its first presi- in the late 1940s. The company had confidence dent. It is in keeping with this tradition that based on previous experience—including the Thompson and Standingford could claim in very recent experiences of the war—that they their report that it was almost a duty for Lyons were capable of effectively organizing busi- to get directly involved in the design of a com- nesses remote from their core food business. puter for business applications to solve the crit- ical problems of the nation (see preceding The academic frameworks quotation from the report). Recent research throws some light on the conditions that have to be fulfilled if an enter- Self-sufficiency. The economist Oliver prise is to be successful and capable of sus- Williamson,9,10 suggests that firms have two tained innovation. Does the research and the types of strategies for coping with transaction frameworks developed help explain how the costs. The first, which he calls hierarchies, Lyons of the 1940s and 1950s came to be first involves vertical integration. The firm attempts in the development of computers working on to minimize transaction costs and retain con- business applications?

July–September 2000 21 First Business Computer

Again organizations vary enormously in the degree of solidarity achieved. High Networked Communal Goffee and Jones note that successful organizations can exist at all points in the two Sociability dimensions, including the extremes. They go on to suggest that we can classify organizational culture into four archetypes: networked, Fragmented Mercenary Low communal, fragmented, and mercenary. Each archetype is associated with a position on the sociability and solidarity dimensions. Again, Low High each can be the basis of a successful enterprise, though they suggest that it is possible to devise Solidarity a contingency framework that indicates the Figure 1. Two dimensions, four cultures. best fit between the cultural archetype and the enterprise. Goffee and Jones plot the two dimensions of culture, sociability, and solidarity against the four archetypes in a two by two Goffee and Jones. The most recent study is table, as illustrated in Figure 1. by Goffee and Jones11 and reports on research Does the Lyons of the 1940s and 1950s fit that they are conducting on the relationship into this model? I used the assessment question- between company success and different styles naire provided by Goffee and Jones (see Figure 2) of company culture. They distinguish between and rated Lyons in terms of sociability as mod- two dimensions of culture. The first dimension erately high. (Note: This questionnaire was not is concerned with the degree of sociability administered to a sample of Lyons’ employees achieved by the employees of the organization. from the 1940s and 1950s, but was completed by The second is concerned with the degree of sol- the author and an ex-LEO colleague.) The rather idarity present among the employees. formal, hierarchical structure of the firm was off- By sociability Goffee and Jones mean a cul- set by the paternalistic attitude of the founding ture that regards the organization as a commu- family, reflected by the social activities sponsored nity and encourages sincere friendship among by the company. There were a large number of its employees. Relationships are often more social events organized through a variety of cul- important than measures of performance. tural and sporting clubs. A relatively large num- Hence it is rare to point the finger at a colleague ber of employees took part in social events and who is under performing. Employees tend to often participated with their families. In these share values and a high value is put on achiev- clubs, rank within the company, except at the ing consensus, even if that means engaging in most senior level, played little part. Ability at compromise. Esprit de corps tends to be high; football, or chess, or bridge, or organizing con- teamwork flourishes; and “a spirit of openness certs counted for more and led to esteem among to new ideas, and freedom to accept out-of-the- the peer group. There was undoubtedly a Lyons box thinking” prevails. There is a wide range of esprit de corps. Sometimes newcomers found it sociability in organizations, from enterprises difficult to enter established networks of friend- with few social contacts and few good rela- ship. This is in keeping with Goffee and Jones’ tionships to enterprises that offer a well-bal- findings that report the formation of cliques in anced community or family. some of the organizations that rated high on the Solidarity, in contrast, implies shared objec- scale of sociability. tives rather than shared values. “Its relationships In terms of the other dimension, solidarity, are based on common tasks, mutual interests, or using the questionnaire Lyons scored a some- shared goals that will benefit all involved par- what higher rating then on the sociability scale. ties.” The organization will respond quickly to The company introduced management by outside threats, and low-level performers are not objectives as a way of ensuring that all employ- tolerated. Members regard achieving consensus ees had a definition of what was expected of as having a low value and think it works against them and, indeed, participated in the definition achieving high performance targets and could of the roles and the way performance was meas- even help competitors. A good example is where ured. Company solidarity was reflected in the in an effort to achieve consensus, difficult deci- arrogance with which members of the Lyons sions are delayed, permitting a sharper com- staff regarded competitors. There was a general petitor to take the lead. Employees are clear feeling that Lyons the company and Lyons about job definitions and organizational roles. employees were an elite.

22 IEEE Annals of the History of Computing As Goffee and Jones point out many managers What is your Organization’s Culture assume that the top right hand quadrant of Figure 1, To assess your organization’s level of sociability, answer the following ques- scoring high on both the tions: sociability and solidarity dimensions, the commu- Low Medium High nal type of culture, is the 1. People here try to make friends and to most likely to indicate a keep their relationships strong successful company. But their research suggested 2. People here get along well that the communal type of culture tends to be unsta- 3. People in our group often socialise ble and is rarely sustained. outside the office Some of the characteristics of high sociability, for 4. People here really like one another example, the reluctance to “shop” the low-level per- 5. When people leave our group, we stay in touch former, are inherently in 6. People here do favours for others conflict with the high soli- because they like one another darity characteristics of 7. People here often confide in one another putting a high value on about personal matters business objectives. Nevertheless, they did find To assess your organization’s level of solidarity, answer the following in existence organizations questions: characterized as communal and for which that 1. Our group understands and shares type of culture was appro- the same business objectives priate and associated with 2. Work gets done effectively and productively success. Goffee and Jones 3. Our group takes strong action to address suggest that strong, charis- poor performance matic, leadership that 4. Our collective will to win is high believes in the values asso- 5. When opportunities for competitive advantage ciated with sociability but, arise, we move quickly to capitalize on them at the same time, is 6. We share the same strategic goals absolutely clear on the 7. We know who he competition is direction the company is driving and doesn’t permit subpar performance is nec- Figure 2. Goffee and Jones’ organization assessment questionnaire. essary to sustain the communal culture. The Lyons of the 1940s and 1950s fits rea- based on a continuous stream of patent-pro- sonably well into the communal culture quad- tected product innovation. But, although rant, scoring quite well on sociability and product innovation is important in the food somewhat higher on solidarity. Does this help to business, there is little protection against copy- answer the question of how they came to make ing, and more competitive advantages may be the decision to design and build a business com- gained from process innovation both in man- puter to serve their management? ufacturing processes and support service Goffee and Jones found three features com- processes. In Lyons, process innovation was mon to enterprises that could be described as encouraged through suggestion schemes and having communal cultures and were success- more formally via the activities of the systems ful over a long period. research office. A stream of new ideas bubbled They found that successful communal cul- up through the organization and many were tures are associated with organizations in implemented over the years. Perhaps the same which extensive teamwork across functions conditions that encourage product innovation makes innovation possible. They give as an in the pharmaceutical industry stimulated the example a pharmaceutical company (Glaxo kind of process innovation that led Lyons Wellcome) in which competitive success is towards the recognition that computers would

July–September 2000 23 First Business Computer

help to make them more efficient and effective. this complexity. Perhaps Lyons survived the In Lyons, the stream of ideas for improvement complexities of the postwar period by having a could only be effectively implemented if the communal culture. More importantly it is a necessary teamwork was deeply embedded in culture that enabled the company to behave in the culture. a dynamic manner. Again the confidence to take on new types What Goffee and Jones’ research point to is of work like designing a computer inside the that culture provides the environment in organization rather than going outside to spe- which success or failure is played out. Certain cialist contractors stems from the same com- types of culture facilitate or enable organiza- munal characteristics. tional behavior. The communal culture Lyons Goffee and Jones’ second feature of a suc- appeared to have in the 1940s and 1950s would cessful communal culture is that there are real lead us to expect a company that encouraged synergies among organizational subunits and and welcomed innovation and was confident real opportunities for learning. Again these and capable of taking the long view and make appear to be important characteristics when it long-term strategic judgments. It appears that comes to disseminating and implementing the it was a company capable of learning and range of innovations discussed previously. therefore absorbing change and innovation. Lyons encouraged the synergy and learning by Goffee and Jones’ research helps us place Lyons widely broadcasting throughout the organiza- in its cultural context and to understand how tion what the company was attempting in its the particular strengths of that culture could LEO venture. Although the rapid growth of the lead to the decisions they made. LEO team necessitated recruiting, mainly of new graduates from the outside, many of the John Kay. The second framework is derived early LEO team members came from a variety from John Kay.6 His research into what factors of job functions inside Lyons. This, too, facili- underlie corporate success found that four dis- tated mutual learning and synergy. tinctive capabilities are associated with the suc- The third finding of Goffee and Jones sug- cessful enterprise. Kay notes that particular gests that a communal culture is important in architectures, good reputation, a flow of inno- organizations that have to take a long view— vation, and the presence of strategic assets are where strategies are long-term rather than characteristics associated with successful enter- short-term. Clearly the Lyons decision to build prises. Any one of these can help to distinguish and apply a computer looks like a long-term an enterprise and lead to its success. Although, strategy, and thus supports the notion that a usually more than one of the capabilities are communal culture would enable long-term present, and the presence of one can lead to views to be taken. However, it is worth noting others. Thus an appropriate architecture can that the Lyons management was somewhat help the business become a successful innova- optimistic both on how long it would take to tor, which in turn can ensure a sustained and a get the computer built and online and the time high reputation. it would take to pay off the investment. Early By architecture Kay means the set of formal estimates suggested that a time period of about and informal relations that exist in the enter- two years for building the computer and a fur- prise. Kay describes architecture as “...a network ther two years to recover the costs would be of relational contracts within, or around, the needed. Perhaps a planning horizon of four firm.” As noted previously, Lyons had devel- years might be regarded as taking a long view oped a management style—really an architec- in the food industry. There is no doubt that the ture in Kay’s sense—that provided clear Lyons management looked at the possibility of information channels between operational and computers as one that influenced not only the management levels unblocked by problems of long-term strategy of the company itself, but as hierarchy and its associated bureaucracy. This something that might affect the whole direc- had three effects. It brought operational man- tion of business practice in the UK. agement and their supervisors into close con- Goffee and Jones note that organizations tact with senior level management, resulting in operating in a dynamic and complex business mutual trust and esteem and hence, effective environment benefited from having a commu- teamwork despite the existence of a notional nal culture. In the immediate postwar period hierarchy. Next, it made quick decision making almost all businesses were faced with the tran- possible at the senior level. Finally, it enabled sition from working on a wartime footing to new ideas bubbling up from the lower levels to that of peacetime. In the food industry the con- be brought to the attention of senior manage- tinuing problems of food rationing increased ment and acted upon speedily.

24 IEEE Annals of the History of Computing Reviewing the case study it is clear that Lyons in the period under discussion had been a company in which a flow of innovations Lyons appears to have had (both process innovations concerned primarily with administrative procedures in the offices, the architecture to welcome and technical processes in the manufacturing units) and product innovations were facilitat- and accept process ed by the distinct architecture developed by the company. Later all three consequences of the innovations of a quite architecture helped to start the LEO project. The mutual trust engendered by the manage- radical nature. ment style and architecture enabled the Lyons board to respond positively and quickly to the report produced by Thompson and Standing- ford without the delays (and obfuscation) gen- erated by some other architectures. argues that it requires other distinct capabilities, The second distinctive capability identified by notably architecture, to enable innovation to Kay is reputation. Kay illustrates the role of rep- translate into business success. Kay’s notion of utation in sustaining company success. A good architecture comes close to the concept of cul- illustration from today (not from Kay) of the role ture as defined by Goffee and Jones. played by reputation is the Virgin organization Lyons appears to have had the architecture and its charismatic founder Richard Branson. characterized by their communal culture to Branson made his reputation early by founding a welcome and accept process innovations of a successful record business. This reputation has quite radical nature. Whether they had the carried over to the extent that the market capability to reap the benefits of all the inno- appears willing to provide support for Branson vations and, in particular, their pioneering whenever he branches into businesses that efforts in adapting information technology to appear to be dominated by established corpora- the business, requires analysis of the later stages tions, such as Coca Cola, British Airways, or the of the initiation, dissemination, and imple- major financial services companies. mentation of the technology. The fact that Lyons had established a reputa- The final distinct capability identified by Kay tion of “can do” in catering and food manufac- is the availability of strategic assets. Strategic turing and in World War II in making munitions assets to Kay comprise features such as the pos- made it more believable that it could also make sibility of controlling the market via a monop- a success of computers. Management had the oly position. Perhaps the LEO venture was made confidence that they could make computers possible by the existence of a supply-side vacu- suitable for business needs, where they also um. Lyons’ strategic asset was the lack of a viable believed that more conventional electronic and supplier in the UK just at the time that business machine companies would be much Thompson and Standingford produced their slower off the mark. Lyons’ management had a report. Another strategic asset was the availabil- belief in their company reputation. But to suc- ity within the Lyons organization of employees ceed they needed others to share that belief. of the caliber of Simmons, Thompson, Would Cambridge university have been willing Standingford, and Caminer. Without this very to collaborate with Lyons in building LEO as a distinctive asset it is doubtful whether Lyons successor to EDSAC if they did not share the would have initiated the LEO project and built confidence in the Lyons reputation? Subse- the world’s first business computer. quently, despite some skepticism, Lyons’ repu- A feature of Lyons at that time was that they tation and in particular the reputation of certain had developed strength in all four of Kay’s dis- individuals, such as John Simmons, played a tinct capabilities, and that they reinforced each major role in the successful launch of LEO Com- other. The right architecture led to innovation. puters as a vendor of business computers. Successful innovations improved reputation. The third distinctive capability in the Kay Lyons had become an acknowledged leader in framework is innovation. He notes that it is innovation for administrative efficiency. The often not enough for an enterprise to have the strategic asset of the lack of supply-side com- ability to make innovations. Too often the petition was perhaps fortuitous, but having the innovator does not reap the benefit of the inno- right people on board was a consequence both vation or cannot sustain the advantages which of reputation —the best people wanted to work the innovations were expected to convey. He for Lyons—and of the architecture.

July–September 2000 25 First Business Computer

Conclusion 7. J.R.M. Simmons, LEO and the Managers, All three strands in the analysis—that based MacDonald, London,1962. on a pragmatic analysis of the case history and 8. T.R. Thompson and O. Standingford, Report on those based on the academic frameworks—con- Visit to USA, archives of J. Lyons and Co., tribute to our retrospective understanding of May/June 1947. how Lyons came to initiate such an unlikely 9. O.E. Williamson, Markets and Hierarchies: Analysis venture as to invent the business computer. and Antitrust Implications, The Free Press, New Retrospective analysis is always easier than pre- York, 1975. diction. Do the three bases of analysis used in 10. O.E. Williamson, Economic Organisation, Wheat- the article, individually or in combination, sheaf Books, 1986. enable us to predict which enterprises will lead 11. R. Goffee and G. Jones, “What Holds the Modern (successfully) in innovation and which are Company Together,” Harvard Business Review, more likely to be followers? These bases may Vol. 74, No. 6, Nov./Dec., 1996, pp. 133-148. help us identify enterprises that have in place 12. R.L. Daft, “A Dual Core Model of Organizational the conditions for making successful and sus- Innovation,” Academy of Management J., Vol. 21, tained innovation possible. They may also help No. 2, pp. 193-210, 1978. identify the businesses that are unlikely to take 13. R.L. Daft, “Bureaucratic versus Nonbureaucratic up a leadership role via innovation. However, Structure and the Process of Innovation and no single framework provides certain answers. Change,” Research in the Sociology of Organiza- Used in combination they may provide greater tions 1, S.B. Bacharach, ed., JAI Press, Greenwich, powers of analysis. Conn., 1982, pp. 129-166. Innovation theory itself has produced a 14. E.M. Rogers, Diffusion of Innovation, 3rd Ed., Free number of theoretical models and frameworks. Press, New York, 1983. Many of these emphasize the diffusion and 15. F. Damanpour and W.M. Evan, “Organizational adaption phases of the innovation phenome- Innovation and Performance: The Problem of non and are less relevant to the earlier inven- Organizational Lag,” Administrative Science Quar- tion and initiation phases. In the later stages of terly, Vol. 29, No. 3, 1984, pp. 392-409. this study, analyzing the adoption by Lyons of 16. E.B. Swanson, “Information Systems Innovation the LEO computer and the diffusion of the Among Organizations,” Management Science, technology first through Lyons and then many Vol. 40, No. 9, 1994, pp. 1069-1092. other companies, more use is made of theories 17. R.W. Zmud, “Diffusion of Modern Software Prac- of technology transfer12-15 including those tices: Influence of Centralization and Formaliza- which are specifically tailored to the special tion,” Management Science, Vol. 28, No. 12, case of information technology and informa- 1982, pp. 1421-1431. tion systems.16-18 18. R.W. Zmud, “An Examination of ‘Push-Pull’ Theo- ry Applied to Process Innovation in Knowledge Work,” Management Science, Vol. 30, No. 6, References 1984, pp. 727-738. 1. D.T. Caminer et al., LEO—The Incredible Story of the World’s First Business Computer, McGraw-Hill, New York, 1998. (Note that this book is the revised US edition of Caminer, User-Driven Inno- Frank Land is currently a visit- vation, see reference 4.) ing professor of information 2. J. Hendry, “The Teashop Computer Manufactur- management at the London er: J. Lyons,” Business History, Vol. XXIX, No. 1, School of Economics and at the Jan. 1987, pp. 73-102. University of Bath. He earned a 3. P.J. Bird, LEO: The First Business Computer, Hasler BS in economics from the Publishing, Wokingham, UK, 1994. London School of Economics 4. D.T. Caminer et al., User-Driven Innovation: The in 1950. Land was employed at J. Lyons and Co. and World’s First Business Computer, McGraw-Hill, joined the LEO team in 1953 and worked with LEO Maidenhead, UK, 1996. until 1967. Recently, Land and colleagues who 5. J.B.B. Aris, “Inventing Systems Engineering,” worked at LEO have published a book LEO—The Proc. Kiev Symp. Computers in Europe: Past, Incredible Story of the World’s First Business Computer. Present, and Future, Int’l Charity Foundation for Land has spent most of his career working with com- History and Development of Computer Science puters and information systems. and Technology, 1998, pp. 33-46. 6. J. Kay, Foundations of Corporate Success: How Busi- Readers can contact the author at Strowford ness Strategies Add Value, Oxford University Press, Barton, Harford, Ivybridge, Devon, PL21 OJD, United Oxford, UK, 1993. Kingdom; fl[email protected].

26 IEEE Annals of the History of Computing