ÐÓ°ÉÔ­´´

In the familiar idea of evolution set out by Charles Darwin, when organisms compete for the same food supply or space in an environment, it is the ‘fittest’ that prevail and survive in the long term, while those less fitted decline. Mutations that confer advantages in the struggle tend to be reproduced and become widespread. The same idea can be applied to everyday technologies: consumers are the environment, and their money is the food that technologies live on and compete for.

As we will see, such a technique is eventually helpful in classifying technologies in ways which emphasise essential similarities. A conventional grouping might say, for example, that both torch batteries and nuclear warheads should be grouped under ‘Energy’ because they both generate it. But try teaching a child that and it quickly becomes clear that there are more differences than similarities, including size, reaction method, speed of energy release and cost. Better to view it all from a neo-Darwinian stance.

Think of music reproduction. The wind-up gramophone of my grandparents’ generation is extinct. In three human generations, five sorts of listening hardware – Edison’s cylinders, brittle 78 rpm records, vinyl LPs and singles, cassettes, CDs – have arrived and be- come pre-eminent; yet already two have gone. This list does not even include total failures. Eight-track tape and quadraphonic sound died out because they were ill-adapted to capture consumers’ money. The heyday of cylinders was ended by discs, which were better adapted for feeding off the world of the consumer.

Or take another class of everyday gadgets: razors. When I started shaving I used a double-edged blade, which was inserted into a robust metal holder. By 1965 this model was an established classic; my grandfather had used almost the same model. Then shaving technology began changing rapidly. Model after model was proffered, each needing a new kind of blade. There was the ribbon, consisting of a long coiled metal strip, which was wound on section by section as it became blunt. That was followed by the disposable, a single blade moulded into a plastic cassette. Then came the ‘twin blade’ in a similar cassette; later versions also swivelled. The most recent type has spring mounts for the double blades. The latest blades are engineering marvels that have made cuts and nicks a rarity. Unlike my father, I do not need a styptic pencil in my shaving kit. So the spring-mounted version survives, and dominates, because its adaptations let it feed off consumers’ wallets.

Several scholars – including George Basalla, professor of history at the University of Delaware – have likened technological change to Darwinian evolution. The inherent problem is the difficulty of defining how ‘fitness’ is established over a century or more, when a host of technologies may rise and fall. But narrowing the focus to a single artefact or pair of artefacts, over a period of a decade or so, allows the ‘evolutionary process’ to be understood by analogy with the established science of ecology.

Growing an example

Ecologists studying the struggles for dominance between two species have been able to extract the underlying mathematics of growth and decline. For example, in the 1930s George Gause, at Moscow’s Zoological Museum, studied the competition between a traditional brewer’s yeast and one used in Ukraine to make a milk drink called kefir. He first grew the two yeasts in isolation, measuring the change in population density with time. Then he grew them together. The brewer’s yeast is tolerant to the alcohol that it produces as it grows; the kefir yeast is less so. In a mixture, this gave the brewer’s yeast an increasing advantage as fermentation proceeded, and it outgrew its competitor. Using a mathematical treatment developed by Alfred Lotka and Vito Volterra – who worked separately, but came up with the same mathematics of population growth at the same time in the 1920s – Gause obtained a good fit to his experimental data (see Figure 1). Such a ‘micro’ approach provides a building block making it possible to construct the enormously complex edifice of the whole of technology. But how is fitness established in individual battles for survival? In biology, when an established species comes into combat with a newcomer over access to food, the newcomer will not succeed unless it performs at least as well as the incumbent. In most cases, it must perform significantly better. The same applies for technology, as has been established by recent research at McMaster University in Hamilton, Ontario. In a 1990 study called NewProd, Bob Cooper and Elko Kleinschmidt (who are professors respectively in industrial marketing and technology management, and in marketing and international business) looked in detail at about 200 new products, of which about half met their financial goals.FIG-mg18594401.GIF

Cooper and Kleinschmidt found nine factors associated with success, and from this they have established a useful computer database and prediction methodology. In the model, as in the study, the most important predictor of success is ‘a superior product which offers unique benefits to the user’. For example, aluminium provided a non-rusting easily-opened lid for drinks cans, unlike steel. Consequently it superseded steel-based predecessors for lids and later the can bodies too (see Figure 2). Similarly, telegraph offices have disappeared and the number of telephone calls in the US each day is now twice the number of telegrams sent each year by the end of the Second World War, when they were at the height of their popularity (see Figure 3).FIG-mg18594403.GIF

The model developed by Cooper and Kleinschmidt is sophisticated enough to show whether a combination of secondary factors can compensate for a product’s inferiority. They have quantified what Richard Foster of the McKinsey consulting firm called ‘the attacker’s advantage’ in his book Innovation, published in 1986. Foster’s semi-quantitative graphical treatment, describing the tactics and strategies of attack and defence between pairs of competing artefacts, was a considerable advance over earlier qualitative thinking. The insights of Cooper and Kleinschmidt, combined with an analysis adapted from mathematical ecology, bring a quantitative appreciation of technological change within reach. In the graphs of Figures 2 and 3, a single attacker’s advantage parameter, A, is sufficient to characterise the battle outcome.

But pair-based competition does not describe the development of all technologies: not all new artefacts have to battle for existence against an incumbent one. Advancing technology often creates previously nonexistent demands from consumers, or satisfies old needs in a vastly expanded way. The electric telegraph was something quite new: an instant letter. Microwave ovens have supplemented conventional ovens and led to new applications, especially in fast food. The motor car did more than simply replace the horse-drawn carriage as it expanded the whole concept of personal transportation in far-reaching ways. Particularly in the US, it has led, for example, to the creation of rich suburbs and deprived inner cities, vacated by the people who previously chose to live there because it was conveniently close to their work. Now that travel is comparatively quick, the well-to-do have moved out to pleasant but expensive suburbs, so it is the poor who live near to centres of work because they cannot afford to buy houses further away.

Perfect placement

Products which actually create their own markets are called placement artefacts. Their growth pattern is more like that of a pair of fruit flies multiplying to fill an initially empty bottle than the growth of competing yeasts. Again, this is a powerful ecological parallel, and its mathematics are simpler than that for attackers and incumbents .

While new products often excite people’s interest, the great majority are simply improvements over existing ones. These so-called replacement artefacts can be divided into two classes, depending on the degree of improvement. The telephone replaced the telegraph, but it was a very substantial improvement: the ability to transmit the human voice led to the wiring of individual homes instead of central offices. In this way a communal product came to be replaced by one that was more personal, so expanding its market into individual homes. The ballpoint pen is another replacement product, though the shift was in the opposite direction from the telegraph: the cherished, personal fountain pen gave way to the ubiquitous ballpoint (see Figure 4). But in both cases the improvements expanded the markets considerably, a key property of this category of replacement artefact.FIG-mg18594404.GIF

The other category is characterised by substitution, one for one, in which the market does not expand. Most new brands fit this description, but there are other examples too. In the US, unleaded petrol is completing its substitution of leaded petrol. Generally, substitution seems to be characterised by zero attacker’s advantage (see Figure 5). Often the substituting product offers nothing new but does remove some disadvantage of the old one: the exhaust gases from cars running on leaded petrol contain lead, carbon monoxide and nitrogen oxides, which all contribute to environmental poisoning. Unleaded petrol causes no lead pollution and can be used by cars fitted with catalytic converters, which reduce the other emissions. The growth in the sales of placement artefacts and the rise and fall of replacement and substitution products can all be described quantitatively in ways that parallel the behaviour of simple biological entities such as yeasts or fruit flies.

But why has the pace of technological change accelerated so rapidly in the 20th century? And why do new biological species arise only after the passage of millennia, while technological products have a life usually measured in decades. The most important distinction is that the gene pools of living matter are unique to their species. Yeast genes do not combine with those of fruit flies. In contrast, the ‘genetic materials’ of technology are based in knowledge and experience, which are openly available through libraries, conversation and especially in patents. These technological ‘genes’ combine in four basic ways to create new artefacts. Using a biological metaphor, these are hybrids, mutants, recombinants and metamorphs.

A hybrid is just an organism that contains characteristics of both parents; cross-bred dogs are a common example. A mutant is the result of small, spontaneous genetic change occurring from one generation to the next, of which physical deformities caused by toxins or radiation are obvious examples. A recombinant results from a particularly powerful ‘reshuffle’ of genetic material, producing a sum that is greater than its two parts: genetically manipulated organisms such as bacteria that can produce insulin after DNA splicing are the best-known embodiment. A metamorph is the result of a mutation or recombination that causes a dramatic, visible, physical change in the new generation.

Applying these criteria to technologies, brands of toilet soap can be viewed as hybrids: they are the product of similar technologies, simply combined. Anything patented with a narrow claim is a mutant as it is distinct in some clear way from its parent technologies while retaining most of their characteristics. But if the patent claim is quite broad the artefact is a recombinant: the change from parent to offspring technologies is greater. Finally, metamorphs arise when disparate ideas are brought together – such as writing and playing the piano. Combine them, and you get the first typewriter, which actually used keys from a piano.

Science has vastly increased the number of possible combinations by unleashing huge quantities of the genetic material of technology – the ideas and knowledge that can be embodied in new products. A second role for science has been to define and expand what is possible – the limits of achievement. Limits serve as achievable targets which challenge and stimulate people’s will to reach them. They also define the boundaries within which technologies may evolve. In 1970 Charles Frank, emeritus professor of physics at the University of Bristol, argued theoretically that polyethylene could be made in a form three hundred times stiffer than the familiar material used to make, say, Tupperware. Today the Dutch company DSM, based in Heerlen, markets a fibre called Dyneema with a Young’s modulus close to Frank’s theoretical value.

Science’s third role is in the application of the scientific method to the development of products. Until this became widespread, technological progress was the province of artisans, and was very much a hit-or-miss affair lacking the methodology needed to design experiments and maximise learning from their results.

The power of the scientific method is that the lessons learnt from error greatly increase the chance of success in the next trial. The reason for this century’s rapid advances in technology and the production of artefacts is a combination of new elements, efficient experimentation and human striving to reach towards ever-expanding boundaries. We have become better at creating the ‘genetic’ mixtures of technologies.

A close shave with technology

When the Gillette company started investigating the applied science of shaving in the 1950s, it was reacting to the putative threat of a chemical cream that would dissolve beards. Though that threat never materialised, Gillette came to understand the scientific basis of blade friction and the mechanical properties of hair. This knowledge set new limits to be overcome by the technology of shaving, which eventually produced the coated, spring-mounted double blades that I use today.

Such technology is just one of thousands which affect people’s daily lives. Just as concepts developed in biology can explain how artefacts arise, the classification of biological species can be adapted to categorise artefacts. Such items are built to meet the basic needs of living; if you classify those needs, you classify technology. Thus it is possible to identify seven ‘kingdoms’: communication, defence, health, packaging, raw materials, shelter and transportation.

All artefacts can be classified into one of these seven, and then further into sub-kingdoms. For instance, health includes the sub-kingdoms of sanitation and medicine. The razor, a cutting product for personal cleanliness, belongs in sanitation. The cassette player resides in the kingdom of communication, sub-kingdom audible. The fountain pen lies in the same kingdom, in the successive sub-classes visual, written, hand. Classifying technologies into kingdoms helps in teaching about them. The similarities and differences become clearer. Nuclear warheads belong to defence, sub – kingdom weapons. Torch batteries are raw materials, for generating electricity. The same? Not at all.

None of this means that every new technology will be perfectly adapted to live off its consumers. Even though product developers know – or think they do – the attributes necessary for success, 70 per cent of the candidates they churn out fail. Of course, that does compare quite well with life, where only the tiniest minority of mutants survive.

Further reading: Innovation by Richard Foster, Summit Books, 1986. Modeling Nature: Episodes in the history of population ecology by Sharon E. Kingsland, Chicago University Press, 1985.

Chris Farrell is a director in scientific affairs at Baxter Healthcare, Illinois.

* * *

1: An attacker’s advantage

The Lotka-Volterra method provides an iterative means of finding the sales of an incumbent artefact Sd for each year that it defends itself against an attacker. If the defender’s growth unhampered by the attacker is known (for example, for a technology like fountain pens a typical assumption might be that everyone starts without one but will eventually end up owning one), it will be possible to determine the values of the coefficients alpha and lambda.

In combination with the values of the sales growth of the attacker, Sa, the difference equation is used to find the best fit to an attacker’s advantage parameter A. The larger A is, the greater the advantage, and the more quickly Sd will fall as the attacker overwhelms the defender.

* * *

2: The shock of the new: planes, trains and automobiles

For substantially original artefacts (or organisms) expanding into virgin territory, no ‘attacker’s advantage’ is necessary and the Lotka-Volterra equation is simplified (right).

This is a form of what is known as the ‘logistic’ equation and was applied in ecology to US population growth by Raymond Pearl in the 1920s. Working at Johns Hopkins University in Baltimore, he was a mentor to, among others, Alfred Lotka and Vladimir Alpatov. After two years, Alpatov returned to Russia, where George Gause was subsequently one of his bright young students.

Another way of viewing the equation is to say that sales in a year St are related to the final asymptotic limit S infinity by two parameters a and b. The iterative formula (see left) then describes the value of S at a time t, and so how S changes over time. The fit of actual sales data to the logistic model is quite a good one for many actual cases such as the food tin sales, (see Figure 3), where S infinity = 29.0, a = 0.8 and b = 0.206.