Within their relatively small research budgets, vendors are understandably reluctant to invest in projects from which they are unlikely to reap many or all of the benefits. Economists call this the “appropriability ” problem: because good ideas diffuse rapidly and can be only partially protected by patents, individual firms cannot be assured of reaping (or appropriating) the benefits of their investments. Practical companies therefore tend to underinvest in the generation of new knowledge and technologies (see Box 2.3).
But IT firms face additional hurdles that make the problem of funding research especially acute. First, the IT industry is known for a rapid pace of innovation, which by all reports has accelerated in recent years. With product cycles as short as 6 to 9 months in many areas of IT, companies must pour resources into product development or risk being quickly left behind. The future is harder for research managers to predict than it was in the 1980s, a shift that increases the perceived risk for companies that invest too much in a particular vision of the future. IT companies that commit resources to projects that extend more than 3 years or so can find themselves abandoned by an unexpected direction of the industry. In such an environment, long-term research is risky unless broadly distributed across a portfolio —as prescribed by the diversity principle explained at the beginning of this chapter.
Other disincentives may be further reducing interest in IT research. One is the so-called network effect: the value of a networked application grows with the number of its users.21 Because of this effect, the market can lock in popular applications, which quickly become difficult to displace. The value of prototype applications in such an environment is small. For example, there is little incentive to do research on technically superior alternatives to common standards such as TCP/IP, Microsoft Windows, or the Intel microprocessor architecture; the rewards are more obvious for products that leverage these de facto standards.22 In this environment, there is less innovation in the form of fundamental improvements, which would challenge the dominant technologies; instead, innovation tends to be seen in new products and services that cleverly adapt these technologies to new market needs.23 For example, the Internet's basic protocol, IPv4, which provides only about 4.3 billion unique addresses, may not provide a large enough address space to meet future demands as the Internet grows. A replacement, IPv6, which is generally thought to not only provide a vastly larger address space but other technically superior features, was developed in anticipation of this need, but because of the high cost of getting everyone to switch protocols it has thus far failed to catch on in the marketplace. To surmount the switching problem, a variety of coexistence strategies are being pursued in the Internet community, but none have yet caught on with a large