This article was published by Business Today's online edition in November 1999. It was written at the height of the internet boom, when IT companies were highly regarded by investors and the media. I was fascinated by the unusual economics enjoyed by internet firms and software product firms, such as network effect and increasing returns to scale. However, I felt it was unfortunate that newspaper/magazine articles wrongly clubbed all IT companies together - both internet and software product firms that had superior economics, as well as IT services outsourcing companies that didn't have such economic advantages (though they did benefit from labor arbitrage). This article tries to explain the unusual economic advantages enjoyed by internet and software product firms.
PDF copy:
http://f1.grp.yahoofs.com/v1/gCGrSUPY7EJgW_lyG6uM0ZbjFfROE7OidgrVhgesnZwzJMwgQKwwy-rBJoMV4h_Gkn4DXJWCXrCa6NuyY-Ka-A/Crazy%20Economics%20of%20Infotech%20Business%2C%20BT%2C%20November%201999.pdf
--------------------------
http://www.india-today.com/btoday/19991122/exclusives.html
Crazy Economics of Infotech Business
By U. Parthasarathy
In 1994 Marc Andreessen and his team created the Netscape browser-and put it on the Net for everyone to download for free. Instead of being ruined, Netscape's subsequent share offering fetched $2 billion. Sun Microsystems created the Java programme language in 1995. When Sun made Java's source code freely available, its shares rose $4.5 billion. In 1983, Apple launched a computer with mouse-and-icon controls, with Windows-based PCs appeared in 1987. In ten years Windows had a monopoly, and Apple had almost vanished. When Java threatened to kill its monopoly in operating system Microsoft retaliated by licensing Java from Sun. In 1997, Sun went on the offensive, supporting Microsoft's Active X technology.
Crazy things seem to happen in the world of infotech, and our traditional ideas of business fail when we try to make sense of them. The success of Microsoft and others cannot be explained in terms of product superiority, marketing wizardry or leadership ability. However well managed, no company ever gets 90% market share that way. Recent theories suggest that infotech businesses exhibit some unusual economic principles not seen in other industries. These principles-network effect, switching costs, increasing returns, and so on-mean that the rules of competition are different in infotech, and so are the ingredients of success. This article takes a look at some of these strange principles, what they mean for infotech companies.
NETWORK EFFECT. Among the most powerful of these economic forces is network effect, which means the power of a network increases with the number of its members. For instance, Netscape's web servers enabled companies to display pictures on their web pages for first time, but to access these customers had to have Netscape's own browser. By giving their browsers free to millions of customers, Netscape compelled companies to buy their servers. And as more companies used Netscape servers, customers had even more incentive to download Netscape's free browser. Netscape's 10 million browser users were de-facto salesmen for its servers, and that accounted for its stunning market capitalisation.
The same logic prompted Progressive Networks in 1995 to let people freely download RealAudio player, their software for playing audio files. The idea was that if more customers have RealAudio players on their PCs, companies will be forced to use RealAudio's servers to add voice and music to their websites. And that is where Progressive Networks make their money.
Network effects are not restricted to computer firms. Consider the battle between Panasonic's VHS format and Sony's Beta format in video cassette players. Beta was supposedly the superior format, but Panasonic managed to push its VHS players overtake Beta initially. Faced with greater demand for VHS cassettes than for Beta, movie makers released more cassettes in VHS. This prompted more new customers to buy Panasonic VCRs.
Eventually movie-makers stopped releasing Beta cassettes, although Network effects are so powerful because they lead to the formation of standards in an industry, especially where technologies are incompatible. A company which exploits network effects will become a monopoly, even if its products are inferior. In industries where compatibility is not an issue, network effects do not exist. In steel, for instance, even if one manufacturer gets ahead initially, a better rival can always capture the market. Similarly, Honda cannot get a monopoly in cars by giving away free tyres.
Even within infotech, network effects do not apply where common standards exist. Any PC can be connected to any monitor, hence Compaq will not have an advantage selling monitors, nor will Samtron have an advantage selling PCs.
SWITCHING COSTS. The second unusual principle underlying the economics of infotech is switching costs. For most items, a customer will shift to a new product if it offers better value and price than his current product. This is true for watches and detergents, but not for computer software. Since it is difficult to install and learn new software packages, people who are familiar with one package will prefer not to switch to another, unless it is vastly superior.
Switching costs explain the continued success of Microsoft's Word, Excel and PowerPoint packages. When Microsoft launched Word97, customers did not have to learn a new package all over again. Word97 retained the look and feel of earlier versions of Word, contained better features, and could open documents created in the earlier versions because it was so difficult to switch to a different package-like Lotus Amipro-customers had only 2 choices: either upgrade to Word97, or retain their old version of Word.
It was because of switching costs again that Microsoft paid $400 million to buy Hotmail from Sabeer Bhatia. The 30 million users of Hotmail would be reluctant to shift to a competitor, since they would have to notify friends and relatives about a change in the e-mail addresses. Many Hotmail users would probably not shift even if a competitor were to offer to pay them. With millions of captive customers, Microsoft can get a steady stream of advertising revenues, as well as extend Hotmail's e-mail service into more profitable areas.
Microsoft's dominance in PCs comes from this impressive ability to exploit switching cost and network effects. When Netscape put its browser on the Net, Microsoft did the same with its Internet Explorer, and also bundled it with Windows 98. Soon Microsoft's browser users exceeded Netscape's, and network effect favoured Microsoft. Similarly, Microsoft is currently adding new features to Hotmail-like Messenger and Passport-transforming it into a portal site like Yahoo. The pattern repeats in Microsoft's reaction to the Java threat. Since Java can run on any operating system, customers will no longer be forced to buy Windows, and operating systems could become unprofitable commodities Microsoft countered this by licensing Java from Sun. Using Microsoft's ActiveX technologies Java programmes can interface better with the PC. Rather than make a package in pure Java, software developers would then prefer to make it in Java+ActiveX, targeting millions of Windows customers. This is network effect at its best-instead of eroding its monopoly, Java could actually help Microsoft. Traditional explanations of Microsoft's success-technology, marketing, and so on-completely miss the point.
Switching costs are not present in all infotech sectors, however. The use of common standards eliminates switching costs in PC hardware and manufacturers of these commodities barely make profits. Companies making customised software do not enjoy switching costs or network effects, and have much lower profits. Developing custom packages is a commodity business-there are many sellers, products are similar, and competition boils down to price. Indian software exporters are relatively more profitable, but that is because their costs are lower. Many software packages are also commodities-MP3 players Winamp and Xing, and all those packages on CD-ROMs we get free with PC magazines. ERP software is profitable because it involves switching costs. However, network effects are absent, so ERP companies like SAP and BaaN are nowhere as profitable as Microsoft.
ECONOMICS OF SCALE. When marginal cost of a product is very low, the industry tens towards monopoly. A package like Microsoft Office 97 takes millions of dollars to develop, but each CD-ROM copy can be made for a few dollars. By selling millions of copies at hundreds of dollars each, Microsoft makes huge profits. Even if competitors make a similar product, they cannot compete with Microsoft if they do not get the same volumes.
Economies of scale effects are absent in other industries because marginal costs rise as firms produce more. A cement manufacturer's cost per tonne of cement will initially fall as capacity expands, but eventually technical and managerial problems will push up his manufacturing cost. In the first stage the manufacturer has increasing returns to scale. In the second stage, we say diminishing returns have set in. Economic theory says that firms will fix production at the level where marginal costs are minimum. Since marginal costs eventually rise, firms in most industries cannot increase production beyond a certain point. For example, Sony's cost per TV falls when it expands capacity from a thousand sets to a million sets, but if capacity rises from one million to 10 million in the same factory, costs may go up. Were it not for this, we would have had monopolies everywhere-one TV manufacturer, one bank, and so on.
What is different about infotech is that marginal costs are low even at large volume. Microsoft's cost of making a CD-ROM is almost the same whether it sells 1 million copies or 100 copies. For a search engine like Altavista, marginal cost is initially zero, and rises beyond a certain number of queries handled, when it will have to invest in bigger servers and extra telecom bandwidth. But at such volumes, Altavista's advertising revenues will rise faster than its marginal costs, increasing profits further. By contrast, in other industries, growth in volume can actually lead to a fall in marginal revenue. Pepsi might sell 1 billion bottles at Rs 9 each, but to sell 10 billion bottles, it will probably have to slash price to Rs 3-below marginal cost.
ECONOMIES OF SCOPE. This is the cost advantage a firm gets by widening its product range. For rediff.com and other online sellers of books and cassettes, greater customer traffic means better understanding of customer preferences, and more precise customer segmentation (say, women in Chennai aged 25-35 who have bought books on gardening). Rediff can charge higher for such targeted advertisements, and can afford to give customers bigger discounts to boost traffic. Rediff can also exploit its understanding of customers' reading habits to come up with highly targeted offers in other products-like cassettes, movies, restaurants, and holiday packages. The more book customers it has, the more cassettes and movie tickets it can sell, and Rediff has the potential to dominate several industries. Many Net businesses enjoy economies of scope, including online financial services.
INCREASING RETURNS TO SCALE. Together, these forces create increasing returns to scale, where the firm's costs reduce as it increases sales. Bigger firms can then undercut smaller competitors and drive them out of business. This tendency is already evident in many Net businesses-browsers, search engines, portal sites, and so on-where big firms get bigger and small ones get smaller.
Of all the economic principles underlying infotech, none is so bizarre as increasing returns. It is not found in other industries, where rising marginal costs lead to diminishing returns, and no firm gets a monopoly. It has no place in traditional economic theory either, where diminishing returns are taken for granted. Increasing returns demolishes this familiar landscape, and gives rise to strange phenomena like path dependence and lock-in.
PATH DEPENDENCE. When the outcome of an industry's evolution depends on what happened initially, it exhibits path dependence. Panasonic's VHS ended up with a monopoly because it managed to get ahead in the beginning, after which it kept growing bigger and bigger. If Beta had got ahead initially, it would have ended up with the monopoly.
Path dependence does not occur in most other industries. For example, consider the equilibrium price of wheat (where demand and supply are equal) is Rs 10 per kilo. If initially the price had been Rs 20, unsold stocks would force sellers to cut prices, and if had been Rs 5, excess demand would encourage sellers to hike prices. Equilibrium is established at Rs 10, irrespective of the initial condition.
LOCK IN. Path dependence demonstrates the role of accidents in shaping the future of an industry IBM's original choice for the PC operating system was CP/M from Digital Research, not Microsoft. IBM was in a hurry, so when the CP/M deal fell through they contacted Microsoft. Microsoft did not have an operating system, so they bought QDOS from Seat Computer Products, modified it, and gave it to IBM. None of the parties-IBM, Digital Research or Microsoft-realised that this would give Microsoft a monopoly. When an industry is in its infancy, products are divergent, advertisements are chaotic, and customers barely know how to compare products. Some products may get an early lead, often through pure chance. It is not possible to predict which one will succeed, because success does not logically follow from product specifications. But once a product gets ahead, increasing returns rapidly make it a monopoly. Even a new, superior product will not be able to dislodge the winner. This situation where increasing returns push a possibly inferior product into a dominant position is known as 'lock in.' This explains the success of VHS video tapes, DOS operating system, the QWERTY keyboard and so on.
EARLY MOVER ADVANTAGES. Lock in gives a pioneer the chance to define standards, get customers hooked on, create economies of scale, and eventually monopolise the industry. Victory usually goes to the company which launches its product first-even if it has bugs and bloomers. However, to exploit early mover advantages, a new entrant must create network effects and switch costs. Apple virtually created the PC revolution, but frittered away its lead. If it had encouraged others to make the hardware and software restricting itself to the operating system, Apple's Steve Jobs would have been the world's richest man.
DIFFERENT RULES. Our perspective of business is conditioned by traditional notions of perfect competition and diminishing returns. These hold good in most industries around us-cement, TV, retailing, banking-where firms get tiny margins, and monopoly does not happen. What is unusual about infotech is that these rules do not apply.
Since the underlying economics is different, and success factors are different, we can't judge infotech firms the way we judge others. For Sony to get 90 per cent marketshare in TVs would need incredibly good R&D, advertising, and so on. But when Microsoft gets 90 per cent share, it can only be ascribed to luck and the ability to pounce on new opportunities. When Indian software stocks zoom on the Bombay Stock Exchange, it is out of place to talk about corporate governance and knowledge assets. It has more to do with low labour costs and price-based competition. Investors scrambled for Netscape shares not because of irrational exuberance, but because they perceived that Netscape could define standards and monopolise its business.
No comments:
Post a Comment