Internet

“The Internet changes everything”—or so we were told at the height of the Internet craze. According to a prominent Wall Street Journal article titled “Goodbye Supply and Demand,” one of the changes it was claimed to have brought about was a transformation of the basic economic forces at work in the economy. After the fact, we know that the Internet did not change everything, and it certainly did not change the fundamental economic forces that underlie the workings of markets. Many economists did not believe that the Internet, or any mere change in technology, could alter the fundamental laws of markets that have been established over the last three centuries.

The Internet’s main function is to reduce the cost of transmitting information (see information and prices), and that is really all it does. Since the Internet only moves electrical pulses from one location to another, it is capable only of transmitting products that can be digitized—and that comes down to various forms of information (including digitized entertainment).

The fact that we can describe what the Internet does in such a brief sentence does not undercut its value as a technology. Not only does it allow the economy to benefit directly from the decreased costs of information transmittal, but also it allows markets to function more efficiently.

Information in the hands of consumers allows them to better choose their products. Anyone who has tried to purchase a car and has consulted the Edmunds or Autobytel Web sites can testify to the wealth of free information available. Stock trackers can do research more easily and collect real-time quotes, something inconceivable for a typical investor before the Internet. Anyone interested in product or market information can quickly find it with search engines such as Google, although this information has not prevented consumers from sometimes overbidding for items on eBay. In addition, many Web sites allow consumers to help other prospective consumers by providing feedback on products.

By reducing the cost of transmitting information, the Internet has also revolutionized some markets, making some local markets into national markets (as with used car parts from scrap yards) and national markets into international markets. It has also allowed people to stay in almost constant contact with friends using instant messaging, has fostered the creation of virtual communities interested in particular topics, and has allowed people with enough energy and interest to become “broadcasters” of information via Web sites and blogs (Web logs).

But despite these changes in marketplace information, the Internet does not change the underlying market forces. The bursting of the bubble—the deflating of the lofty market capitalizations of firms related in any way to the Internet—was itself the most pedestrian case of time-honored economic forces coming to the fore. The law of demand states that the price the marginal consumer is willing to pay declines as the output increases, holding other things constant. This means that price falls as supply increases. Applied to investment, it means that returns fall as the level of investment increases.

Because expectations for the Internet were so high and so widely held, firms promising to use the Internet in their business plans drew a tremendous amount of investment. Some of these business plans were just so much thin air with no potential for any profit. But even companies whose business plans made sense were unable to translate them into profits when a dozen other firms were investing in the same idea. So much investment money was thrown at “Internet” firms that the rate of return was negative.

Of course, consumers benefit from these new investments even if the investors do not. This is also consistent with simple microeconomics, which teaches that although the price falls as supply increases, the benefit to consumers increases.

Some relatively new economic concepts predating the Internet were often used in discussions of the Internet. These concepts—network effects, winner take all, lock-in, first mover wins—are often misunderstood. The last two on the list, as they are commonly used, are not supported in the economics literature.

Many products and firms associated with the Internet are thought to have an economic property known as “network effects.” Network effects are present when a product becomes more useful to consumers as the number of other people using it increases. For example, the owner of a fax machine benefits from the fact that there are many other people with fax machines.

While some products associated with e-commerce did have network effects, many products did not have even the slightest trace of them. Most Internet retailers, for example, had no network effects to speak of. Take the example of E-toys. Although consumers were unlikely to care about the size of the network—how many other consumers were using E-toys—many pundits assumed that any business conducted over a network had network effects.

When network effects are strong, firms with large networks have an advantage over those with small networks. Everything else being equal, consumers should be willing to pay more to join a large network. This advantage of large over small is referred to as winner take all, even though consumers tend to have differentiated tastes and market shares are almost never 100 percent. This is an advantage of size on the demand side, whereas such advantages had previously been thought most likely to arise on the supply side, when firms had economies of scale.

Could the transformation of business from brick and mortar to Internet-based turn an industry that was not previously winner take all into one that was? For the large number of Internet firms that did not enjoy network effects, the answer had to be no. Even for those that had network effects, it need not be, and generally was not, true that the network effects would be strong enough to convert an otherwise typical firm facing the law of diminishing returns into an increasing-returns firm.

Nevertheless, there can be cases where network effects help create winner-take-all characteristics. The software industry might fit that description, although network effects probably play second fiddle to traditional economies of scale—caused by large fixed costs—in creating winner-take-all conditions. The network effects associated with businesses such as eBay, where consumers value having many choices, might be responsible for winner-take-all results.

Winner-take-all markets do not require that the same winner keep winning time after time. It is possible that there will be swift leadership changes whenever a better challenger enters the market.

The concept of lock-in as it is most commonly used—that is, a strong version of lock-in—suggests otherwise. According to the strong version, the winner not only takes all, but also continues to do so even in the face of a better rival. The theory of just why this would be so has to do with a particular type of coordination problem associated with network effects.

Whenever consumers attempt to determine which brand or type of product to buy in network markets, they need to take account of the strength or size of the competing networks. For example, a consumer in the early 1980s determining whether to buy a VHS or Beta VCR would normally have considered the prospective size of the two networks.

Strong lock-in proposes a tantalizing possibility. Every consumer might know that a better technology exists, but because they all fear that other consumers will not switch, everyone remains with the inferior product. Nevertheless, strong lock-in ignores the possibility that consumers will expect others to pick the better product or that the producer of the better product might engage in practices to convince consumers to switch, such as low pricing or performance guarantees.

If the strong lock-in held, then firms would succeed by getting to market first and largely ignoring relative quality because even with a significantly better product a challenger could not dislodge the incumbent. This is the genesis of the first-mover-wins doctrine. This thinking led to the massive frenzied investments that occurred during the Internet bubble and played an important role in the Microsoft antitrust case.

Although this form of lock-in is theoretically possible, before presuming that it is a serious problem we should check to see if it actually happens.

There is no evidence that strong-form lock-in actually occurs. That explains why Altair, VisiCalc, and Ampex—the first firms to produce PCs, spreadsheets, and VCRs, respectively—are not today the leaders in those markets. And it explains why Apple, Lotus, and Sony, the second-generation leaders, are not the current leaders in those markets.

Some economists claim that real cases of strong lock-in do exist. The two most popular alleged examples are the typewriter keyboard and the VCR. The problem, however, is that these examples are counterfeit.

The keyboard story starts with the claim (without any supporting evidence) that to prevent jamming of the keys, the typewriter mechanics who worked on the original QWERTY (named after the upper lefthand keys) machine in the late 1800s came up with a design to slow typing down. In the 1930s, a professor of ergonomics at the University of Washington, August Dvorak, patented his own keyboard, painstakingly created from a systematic study of keyboard design. Dvorak’s own research claimed that this keyboard design worked much better than the QWERTY design, but lock-in advocates give more weight to a U.S. Navy study that supposedly demonstrated that Professor Dvorak’s design was indeed at least 40 percent faster than the QWERTY design.

This story is flawed in many respects. First, it ignores a heavily publicized General Services Administration (GSA) study comparing the two keyboards, which concluded that Dvorak’s design was not superior to QWERTY. Second, it neglects to mention that the Navy’s chief expert during its tests was Lieutenant Commander August Dvorak, the patent holder of the alternative keyboard, who was identified by the GSA study as the author of the Navy study. When Stephen Margolis and I examined the Navy study, we found several important biases and errors. Finally, the story ignores modern ergonomic studies of the keyboard that are inconsistent with the claims of Dvorak advocates.

The VHS/Beta story, claiming that Beta was the better format but that VHS was locked in, is equally flawed for many reasons, not the least of which is that Beta came first. Further, the developers of Beta and VHS had a patent-sharing agreement and used the same technology, with the two formats differing mainly in the size of the tape. This tape size, which gave VHS a longer playing time, proved crucial.

Various people have claimed that the internal combustion engine was locked in at the turn of the twentieth century, preventing steam or electric engines from taking hold. Yet, as a student of Paul David’s concluded after writing his dissertation on the subject, this claim was false. So far, no examples of strong lock-in have withstood scrutiny.

In our lengthy examination of software markets, Margolis and I found that the product that wins also happens to be as good as or better than the others. A 1999 study I conducted for McKinsey produced similar results. I looked at twenty different markets, ranging from high tech, such as Web portals, to low tech, such as athletic apparel and discount retailers. The results were consistent with those found in software markets. Finally, work by Peter Golder and Gerard Tellis refutes many previous examples of purported wins by first movers.

The Internet will change certain markets. It can play an important role in retailing, particularly when information is of central importance. It will likely become the primary mode of distribution for all things based on information—music, software, books, movies, travel bookings, and so on—although various copyright mechanisms need to be worked out before this becomes true in all cases.

Firms might forgo bricks and mortar if the products they sell have particular characteristics, including, low shipping costs relative to value. Since Internet firms must use delivery companies to get the product to consumers, it is important that the shipping costs not overwhelm the value of the product. The products must be nonperishable goods that can go through a delivery process without losing their value (which was why Internet grocery sales made no sense). Experience goods, which consumers can evaluate only in person, cannot be sold through a virtual experience such as the Internet; nor can goods intended for instant gratification, because shipping time removes the “instant” component of instant gratification. And of course, even brick-and-mortar firms might benefit from using the Web to promote their products or take orders.

Many Web sites have had difficulty supporting themselves. Again, this was partly due to too much investment leading to too many Web sites chasing too few dollars. But there were other miscalculations as well. Although originally many sites expected to charge subscribers for use, this business model was quickly replaced by an advertising-only model adopted from the television industry.

The success over the years of television broadcasters might have provided an appealing exemplar for Web sites, but television has some advantages not shared by Web sites. First, and most important, competition is restricted in the television market since the FCC grants a limited number of licenses, allowing television stations to earn above-normal returns for many years without fear of having entry compete them away, as would happen in a competitive industry.

In addition, television has a superior advertising methodology. The advertisements are embedded in the programming in such a way that one cannot bypass the advertising while consuming the programming (short of averting one’s eyes, leaving the room, or changing channels). The same is not true of most Internet advertising, although popover ads are an attempt to diminish the ease of ignoring the advertising. Further, television advertising is far more engaging than Internet advertising has been and is likely to be for the foreseeable future.

Finally, television has a larger audience—the average household still watches more than seven hours per day, with the typical person watching almost four hours per day. The most recent estimate for Internet usage (2004) is twenty-eight minutes per day per person, with projections of thirty-three minutes per day expected in 2009.1

The size of the relative audiences, compounded by the superior advertising capability of television, allows a simple estimation of the maximum reasonable advertising revenues from the Internet. In Rethinking the Network Economy, I compared the sizes of the two audiences, assuming equivalent advertising capability, and discovered that Internet advertising was already higher than could be reasonably expected. Thus, there was little room for future growth without major increases in Internet usage or more compelling advertising except for search engine advertising.

Web sites hoping to support themselves with their content, therefore, will need to rely increasingly on subscription revenues if they are to survive.