Engines That Move Markets (2nd Ed)

by Alasdair Nairn

  Thus relieved of capital commitments and ongoing services liabilities, Hollerith was in a position to continue his efforts to expand his business. Within a short space of time his equipment had proved its worth with the railroads and a lucrative order was received. Shortly before the end of 1896 his efforts to build a business outside the US also started to bear fruit. After much lobbying, Russia signed up for his equipment with an order worth $67,000 ($5m). Hollerith was particularly astute with this order, insisting on full payment in advance rather than the usual lease terms that he applied in the US. As a consequence he avoided the sovereign risk that was to arise when the Russians started to default on their debts. Hollerith was also astute in recognising the benefits of retaining control over the supply of data cards. In many ways this was similar to the policy of Marconi in leasing equipment with operators. The supply of cards ensured a lucrative ongoing relationship which often exceeded any profits from the supply of the equipment itself.

  Buoyed by the success with the New York Central and the forthcoming Russian order, in December 1896 Hollerith incorporated the Tabulating Machine Company in 1896, thus laying the foundations for the industry giant (IBM) that was to emerge years later. The new company was financially viable and soon sought to obtain greater control of its own destiny by ending the agency contract with the Library Bureau and bringing the manufacture of equipment under its own roof. The company received the contract for the 1900 census from its new head, William Merriam, his predecessor Porter having returned to Britain to incorporate his own version of Hollerith’s endeavour, the British Tabulating Machine Company (later International Computers).

  Unfortunately the relationship with the Census Bureau deteriorated from that point forward and when Merriam, who had left the bureau to become president of the Tabulating Machine Company, began negotiations with his replacement he found that the environment had changed dramatically. The turn of the century had marked a substantial change in sentiment towards ‘big’ business. The trusts were now perceived as enemies of the people, abusing their monopoly position at the expense of the country. This sentiment was not reserved solely for Standard Oil and the railroads. Any large successful company tended to be tarred with the same brush and the Tabulating Machine Company found itself in this category.

  The new superintendent at the Census Bureau believed that the census was being overcharged and that better terms could be achieved. The difficulty was that there was a dearth of competitors to Hollerith’s product and the bureau’s negotiating position was ostensibly weak. Its response was to create a competitor. He sought $40,000 ($2.4m) funding for a research department to create alternative equipment. He also altered the patent that protected Hollerith machinery. In defending itself against this patent infringement, the problem for TMC was that there was no legislative provision for litigating against the government. In any event, suing your largest customer was self-evidently not a prudent move. The litigation that followed was ultimately inconclusive as, by the time the final decision was reached, the 1910 census equipment had been installed.

  For the Census Bureau, the net result of its pre-emptive strike was the creation of a competitor, but at the expense of substandard equipment and higher cost in the short run. The competitor became a private entity when the head of the department set up by North, a Russian immigrant with the anglicised name James Powers, left to form his own company. For TMC there was the satisfaction of the vindication of its equipment, but set against this was the new and real competitive threat of Power’s company. For Hollerith personally, the battle with the government, combined with his ill health, might have contributed to control of the company passing to an entrepreneur named Charles Flint.

  Flint had previous experience in business. Like many businessmen, he sought to obtain the best prices he could for the products he sold. He also recognised that the acquisition of Hollerith’s company would create a powerful consolidated group that could dominate the industry. Under the deal that followed, the ailing Hollerith was paid $1.2m ($65m) for TMC, which allowed him to live out his life in luxury. The new group was called CTR, and the same company that Thomas Watson joined as general manager after NCR. Watson’s background at NCR made him an ideal candidate to take the new company forward. He had seen firsthand the growth in the market for business machines and the importance of incentivised sales techniques and after-sales service. Realising the potential that CTR held, he carefully negotiated himself an employment contract which allowed him a share of the company’s profits as part of his compensation.

  So far as Sloan & Chase, the company created by Powers, was concerned, it was eventually sold to Remington Rand. Remington Rand later became part of Sperry Rand, which in turn ultimately transmuted itself into Unisys. The main lesson was that while nature might abhor a vacuum, customers abhor a monopoly, even if that monopoly delivers the best product. Governments are often the most assiduous fighters against monopoly. The example of Hollerith’s fallout with the Census Bureau is not an isolated one. Governments have repeatedly treated existing statutes and patents as minor obstacles when they conflict with a course of action they wish to pursue.

  The Computing Tabulating Recording Company (later IBM)

  8.2 – IBM: great till the eighties

  Source: IBM annual reports. CRSP, Center for Research in Security Prices, Graduate School of Business, University of Chicago, 2000. (Used with permission. All rights reserved. www.crsp.uchicago.edu.)Commercial and Financial Chronicle. New York Times.

  The most striking element to the progress of IBM is not the steady growth of the early years but the way in which it was completely overshadowed in the period after World War II once the mainframe computer had made it a mainstream company. In the early years of the company, income growth was very strong, averaging somewhere between 15% and 20% per annum, and doing so against a backdrop of a strengthening balance sheet and rising returns on assets and equity. As positive a picture as this was, it bore no relationship to what happened between 1950 and 1970 when the growth path of income turned from strong to a pattern verging on the exponential. Although growth slowed slightly after the initial impact of the introduction of the personal computer, the return IBM earned on invested capital remained in double digits.

  The most interesting aspect of the history of IBM’s returns is the decline in the return on assets, offset by a slight improvement towards the end of the 1990s. The sharp upward movement in the return on equity is marked, rising to levels well in excess of anything earned during the most vibrant periods of the company’s distinguished history. These returns did not flow just from a rapid increase in net income, but from a capital restructuring which saw the debt/equity ratio advance to over 100%. Since 2000 the level of financial gearing has continued to rise and with it the return on equity. Return on assets has not followed the same trajectory, suggesting that profitability is ever more sensitive to financial engineering and vulnerable to a rise in debt costs.

  In February 1924, ten years after the arrival of Watson, CTR changed its name to International Business Machines (IBM), a name it had been using in Canada since 1917. The background was a history of strong growth in revenues and a continued buoyant market for office equipment. IBM was not alone in sharing in the growth of the time. In 1886, William Burroughs had formed the American Arithmometer Company to manufacture and market adding machines. These were targeted mainly at the banking sector where Burroughs had gained his experience. After initial slow progress, sales began to pick up and by the early part of the 1900s the company was selling over 4,000 units a year, with sales doubling every few years. In 1906 the company’s name was changed to reflect its founder and called the Burroughs Adding Machine Company. Adding machines represented a growing business, but the adding machine with the greatest practical impact, and hence the largest sales growth, was the one that assisted with day-to-day transactions, the cash register. National Cash Register (NCR), the company that had grown to service this demand, became the corporate giant of
the time and introduced a range of selling practices which were to become a major influence in shaping the developing industry.

  American Arithmometer/Burroughs

  8.3 – Adding value: the Burroughs way

  Source: American Arithmometer annual reports. Burroughs annual reports. Commercial and Financial Chronicle. New York Times.

  The financial statements of American Arithmometer which sold Burroughs adding machines and later changed its name to Burroughs show clearly the early growth of the industry. The product was a precision-engineered machine and a market-leading product. Sales growth was strong and the firm’s balance sheet was initially adequately capitalised, meaning that recourse to outside financing was not required. Strong growth and product leadership meant that premium prices could be charged, allowing a net income margin that rose from over 20% to nearly 50% at its peak. Margins of such magnitude would not be sustainable in the long run as the key technology was not patent-protected. Nevertheless, there was sufficient protection to allow these margins to be sustained for a ten-year period. As a consequence there was a level of cash flow that more than covered any costs of expansion and still allowed a strong and rising level of dividend payment. Even with the dividend payment, cash built up on the balance sheet, leaving the company in a strong positive net cash position. This is revealed in the closeness between the returns earned on equity and assets. That shareholders’ funds almost equalled total assets reflects the accumulation of profits on the balance sheet rather than any dilution, and hence – as with, for example, Standard Oil – the buildup in value for shareholders would be revealed in the advancement of book value which more than quadrupled over the ten-year period. One would therefore have expected a total return from investing in the company somewhere in the 30–40% range. The annual reports also show that the supportive environment that had allowed such returns to be generated was beginning to fade. The rate of growth in sales was slowing, margins were contracting and as a consequence the return on assets and equity was falling. Absent some new technological shift or additional line of business the expectations for future returns would necessarily be lower than what had been earned in the previous ten years. To the extent that the share price was predicated on a continuation of previous growth and returns, there was a strong chance of disappointment.

  The next wave of innovation

  The economic growth of the industrial world and the profits which accompanied it necessitated a different approach to business management. The requirements for large-scale manipulation and processing of financial and commercial information created a huge demand for mechanical means of simplifying and speeding up the process. Although electricity had entered the equation, it had done so only as an alternative method of powering the mechanical adding equipment. The constraints on expansion of this equipment were therefore based on the capabilities of precision engineering, just as they were in Babbage’s time. The main difference was that there was an eager and growing client base which extended well beyond government into the private sector.

  In the technological sense the need was to utilise the properties of electricity to replace mechanical components. Some of the skills and knowledge required already existed, although these were fairly disparate – spread as they were through a wide range of industries and being applied to related but different applications, from the technology of the radio and television through to power generation and lighting. Most important of all, though, was the role of the government and in particular the needs of the military.

  The machines being manufactured and sold remained largely unchanged in their capabilities. They were effectively simple adding machines, albeit ones that could operate at increasing speeds and with ever larger volumes of information. Many companies sought to establish a foothold in the industry, drawn by its growth and the perceived profit potential, but it remained a relatively small business sector when compared with the steel, chemical and automobile industries. As a new business sector, it also competed for capital with the radio and television companies that by this time were beginning to understand the commercial potential of broadcasting, in addition to the point-to-point transfer of information. Moreover, other companies that had grown rapidly over the previous decades, such as AT&T and GE, also found they could not ignore the potential in the area. New technological applications of electrical power, for example, demanded substantial mathematical modelling in order to predict the outcomes of various combinations of events. Such modelling could take two principal forms: analogue and digital. Analogue modelling at its simplest consisted of building abstract scale models, trying to mirror possible conditions, and from there attempting to extrapolate the likely results. Lord Kelvin (one of the technology giants of the 19th century) had, for example, built an analogue model to help predict tide patterns. The mechanism simulated the effect of the interaction of gravity to allow prediction of tide levels and therefore the plotting of tide charts for harbours. Similarly, when companies such as GE or Westinghouse installed new power networks, they had to work out the effect of linking up different systems. They used massive simulators to observe the range of potential outcomes. The common characteristic of these analogue models was that they were designed for a specific purpose and by definition had limited use.

  Some analogue models were designed in such a manner that they could solve a particular classification of problem, but even these were limited in their general application. In addition, the accuracy of the analogue machines’ calculations was limited by the accuracy of the model, which in turn depended upon the precision of the machinery. It is a testament to the scientific knowledge and expertise which went into their construction that wave machines, for example, continued to be used well after World War II.

  The main point is simply to distinguish these methods of analysis from their numerical or digital counterparts. Where numerical computing differed was that it required no physical representation of the problem being modelled. If the problem could be modelled mathematically, it could be solved numerically – assuming the resources were there to allow the necessary computation to take place. In the 1930s the process of computation remained largely manual, although rather than the tally sticks of the previous century, the ‘human computers’ made use of the new adding machines. Just as better organisation had improved efficiency for Napoleon a century before, so in the 1930s individuals such as Leslie Comrie in Britain revolutionised numerical analysis for astronomical work. The advances achieved by Comrie soon gave way to other uses, most notably ballistics.

  Interest in computing power also spread geographically. Adding impetus to this spread were three factors. Firstly, office equipment companies were increasingly aware of the potential for extending the uses of adding machines and as such had ongoing research projects to this end. Secondly, the scientific community continued to seek practical solutions to produce working models that would provide them with the computational ability they needed to further their theoretical work. Finally, as the 1930s proceeded, the threat of war loomed ever larger, and governments increasingly turned to the business and scientific communities to provide assistance in the mathematically dependent fields of ballistics and code-breaking. In other words, there was a confluence of scientific interest and liberal funding from both the private and public sectors. Overlaid on top of this was national interest, which as had been shown with the radio during World War I, was an irresistible force for breaking down commercial secrecy and spreading information between hitherto separate groups.

  The legacy of Bletchley Park

  In Britain, efforts centred on Bletchley Park, nominally the Post Office Research Station, but in reality the top-secret home of the code-breaking team headed by the mathematician Alan Turing. The task set for Turing was to construct a machine able to decode messages sent by the German Enigma coding equipment. Turing was supplied with an outline of Enigma by a Jewish Polish engineer who had worked in an Enigma assembly plant and who had been smuggled out of Warsaw by the French intelli
gence service. While Turing therefore had an outline of how Enigma worked, he had to work out how to find the ‘cipher key’ which would allow conversion of the encrypted message back into German.

  What Turing developed was the basic algorithms of artificial intelligence, which would allow the shortcuts necessary to narrow down the many billions of possible outcomes to a manageable number which could then be inspected – thus allowing the key to be found. With a model of the Enigma machine, artificial intelligence techniques and a physically huge electromechanical set of processing equipment, the team at Bletchley was able to decipher the German Enigma messages. Technology would not stand still, though, and Turing’s team soon had to contend with an improved German ciphering system three times the size and complexity of the initial version. The first iteration of the British response was the Colossus, which comprised nearly 2,000 vacuum valves. The purpose of the Colossus was to assist in preparing a subset of potential outcomes which could then be processed further to find the underlying message. Arguably, although the British Official Secrets Acts makes the assertion difficult to sustain, Colossus represented the world’s first electronic computer.