By Scott A. Shane (Wharton School Publishing)
From Chapter 1: Selecting the Right Industry:
If you are thinking of starting a new technology company, you can and should examine how favorable different industries are to new firms. A former PhD student of mine, Jon Eckhardt, who is now on the faculty at the University of Wisconsin, compared the proportion of startups in different industries that made the Inc 500, the magazine’s list of the fastest growing young private companies. Looking from 1982 to 2000, Jon found very high levels of variation in this measure. For instance, Jon’s data show that the odds of starting a biotechnology company that made the Inc 500 were 265 times higher than the odds of starting a restaurant that made the Inc 500 and that the odds of starting a computer software company that made the list was 823 times higher than the odds of starting a hotel that made the list. In short, an average entrepreneur starting an average new firm will be more likely to create a high-growth private company or a newly public company in some industries than in others.
So how do we explain these data patterns? Unless most of the talented entrepreneurs are drawn to some industries (e.g., biotechnology) and not others (e.g., hotels), some industries are just better for starting new firms than others. This, of course, means that you need to understand the characteristics of industries that make some of them better for starting firms than others if you want to increase your chances of success. Research has shown that four different dimensions matter: knowledge conditions, demand conditions, industry life cycles, and industry structure.
Certain industries are more favorable to new firms than other industries because of the knowledge conditions underlying the industry. Why? Because some knowledge conditions are easier for new firms to manage, while others require the expertise of established firms. One aspect of knowledge conditions is the level of complexity of the production process. Production in some industries is more complex than production in other industries. For example, the production process in the aerospace industry is more complex than in the paper bag manufacturing business because the number of factors that need to be incorporated into the production process, the level of precision involved in making the products, and the sophistication of the required knowledge are all higher in aerospace than in paper bag manufacturing.
1. Don’t start a business without investigating how favorable the industry is to new firms.
2. Don’t fight the odds. Don’t start a business in an industry that is unfavorable to start-ups.
Industries that involve very complex production processes tend to be unfavorable to new firms. Complex production processes require sophisticated organization structures to coordinate the activities of people engaged in a wide variety of activities. These sophisticated organization structures are easier to implement in established firms that have more specialized labor, larger management teams, and routines for the management of complex operations. Moreover, the knowledge of how to undertake complex activities is often developed through learning by doing over time. Therefore, new firms often have less knowledge of such activities than established companies. Another aspect of knowledge conditions is the amount of new knowledge creation that is required to generate the industry’s products and services. For instance, the pharmaceutical industry relies very heavily on the creation of new knowledge to produce its products and services. Absent basic scientific research, it would be difficult for pharmaceutical researchers to create new drugs. In contrast, the dry cleaning industry does not rely very much on basic scientific research to create its products and services. In fact, dry cleaning services can be offered without very much new knowledge creation at all.
Industries that rely greatly on new knowledge creation, as measured by the proportion of their sales that they devote to research and development (R&D), are less favorable to new firms. New firms perform worse in more knowledge intensive industries because they do not have the internal cash flow to invest in basic research – a situation that creates a larger handicap in R&D intensive industries than in other industries. Moreover, basic research is often very uncertain and can result in the creation of new products or services in very different lines of businesses than those for which it was originally intended. Large, established organizations with economies of scope are more likely to benefit from investing in this type of uncertain research and development than small, new firms. This does not mean you can’t start a successful new company in an R&D intensive industry; many founders of biotechnology companies have done just that. It only means that the R&D intensity is making it harder for you to succeed.
Another dimension of knowledge conditions in an industry that affects the performance of start-ups concerns the codification (the writing down) of knowledge. In some industries, the knowledge necessary to undertake the development of new products and services is readily available in written form. For instance, there are numerous books and articles on computer networking. However, in other industries, this knowledge lies in the heads of experienced personnel who know how to undertake that activity effectively, but who cannot specify in written form the causal mechanisms that lead to performance. Some aspects of circuit design, for example, are known to only a handful of individuals.
Codification of knowledge enhances the performance of new firms because codified knowledge is more easily available to entrepreneurs than is tacit (residing in the heads of a few) knowledge. Because codified knowledge is written down, it is available to entrepreneurs who do not have direct operating experience in the industry. In contrast, tacit knowledge is only available to entrepreneurs with direct operating experience in the industry or who hire those with direct operating experience. As a result, learning curves are less proprietary in industries in which knowledge is codified, allowing new firms to more easily learn what predecessors have learned about an industry and catch up in terms of performance.
Yet another important dimension of knowledge conditions concerns where the innovation that makes new products and services possible is developed. In some industries, such as semiconductors, innovation occurs within the industry itself (sometimes referred to as the “value chain”). Firms and their customers and suppliers generate most innovations. In other industries, such as superconductors, extra-value chain organizations, such as universities, produce much more of the innovation than firms within the value chain.
New firms perform best in industries in which extra-value chain entities produce most of the innovation. Universities and public research institutions are less concerned than firms about keeping valuable knowledge from leaking out to others. In fact, senior managers at companies like Intel take explicit actions, such as having employees sign non-compete agreements and restricting access to research laboratories, to keep valuable knowledge within the organization. In contrast, senior administrators at universities, like MIT, take explicit actions, such as encouraging corporate tours of laboratories and sharing papers at industry-wide meetings, to encourage knowledge transfer out of the organization. Because new firms cannot create much of the knowledge necessary to innovate, industries in which public sector organizations conduct much of the knowledge creation, and knowledge spillovers are relatively large, are those in which new firms perform the best.
Industries also differ in the distribution in proportion of value added that comes from manufacturing and marketing activities, as opposed to product development and innovation. In some industries, such as automobiles, much of the value added comes from manufacturing and marketing, rather than from product development. In other industries, such as software, manufacturing and marketing account for a much smaller proportion of value added, with some companies having virtually no manufacturing or distribution assets. New firms tend to perform poorly in industries in which manufacturing and marketing account for most of the value added. When firms create new products and services, they often need manufacturing and marketing assets to exploit those innovations. As a result, established firms figure out how to market and manufacture effectively and make those activities routine. New firms are at a disadvantage when competing with established firms because they have not yet figured out how to make these activities routine. So their manufacturing and marketing are often quite inefficient.
Moreover, manufacturing and marketing assets are often very expensive – think about the cost of creating an automobile plant – and are difficult to outsource. Manufacturing operations often need specialized assets, such as expensive machines designed to produce specific products, making it very difficult to find someone willing to provide those assets on a contractual basis. As a result, established firms tend to own the necessary manufacturing and marketing assets, making it difficult for new firms to get access to equivalent marketing and manufacturing assets at the time of founding, which handicaps them in their efforts to compete with established firms. Thus, new firms tend to perform more poorly in industries such as automobiles, where the manufacturing and marketing assets are important, than they do in industries such as computer software, where these assets are less important.