Take the case of software products for mass use for example. Creating such a product starts identifying a felt need (say, for a browser once the internet was opened up to the less computer savvy or for a web-publishing programme once the internet went commercial). The persons/firms identifying such a need must work out a strategy of generating the product, by hiring software engineers, at the lowest cost in the shortest possible time. Once out, the effort must be to make the product a proprietary, industry standard. This involves winning a large share of the target consumers, so that the product becomes the industry standard in its area. Once done, the product becomes a revenue generating profit centre. The investment required is the sums involved in setting up the company, in investing in software generation during the gestation period, and in marketing the product once it is out so as to quickly win it a large share of the market. Needless to say, while entry by individuals or small players are not restricted by technology, they could be limited by the lack of seed capital. This is where the venture capitalists enter, betting sums on start-ups which if successful could give them revenues and capital gains that imply enormous returns. 
 
There are, however, two problems here. The first is one of maintaining a monopoly on the idea during the stage when the idea is being translated into a product. The second is that of ensuring that once the product is in the public domain it is not replicated by competitors who win the market before the originator of the idea consolidates. It is here that a feature of 'entrepreneurial technologies' – the easy acquisition and widespread prevalence of the knowledge base needed to generate new products - considered an advantage for small new entrants actually proves a disadvantage.
 
There are two aspects of technology that are crucial in this regard. First, their source. Second, the appropriability of the benefits of a technology. As mentioned earlier, in industries with routinised technologies the source of technology was in significant part the activity of incumbent firms themselves. On the other hand, in the case of entreprenurial technologies the sources were in the public domain. This was where the advantage lay for the small operator. But once a technology is generated based on some expenditure in the form of sunk costs, there must be some way in which the innovator can recoup these costs and earn a profit as incentive to undertake the innovation. In the Schumpeterian world this occurred because of the 'pioneer profits' that the innovator obtained. The lead time required to replicate a technology itself provides the original innovator with a monopoly for a period of time that generates the surplus which warrants innovation.
 
Most often this alone is not enough to warrant innovation and in the software sector lead times can be extremely low, especially if the competitor invests huge sums in software generation, reducing the lead time substantially. It is for this reason that researchers have defended and invoked the benefits of patents and barriers to entry in production, which allow innovators to stave off competition during the period when sunk costs are being recouped. Unfortunately, neither are the status of patents and copyrights in the software area clear ( as illustrated by the failure of Apple to win propietary rights over icons in user interfaces), nor are their barriers to entry into software production.
 
This has had two implications. First, the importance of secrecy in the software business. The 'idea' behind the product must be kept secret right through the development stage, if not competitors can begin rival product developments even before the original product is in the market. A feeble attempt to institutionally guarantee such secrecy is the now infamous 'non-disclosure agreements' which prospective employees, financiers and suppliers are called upon to sign by the innovator who is forced to partially or fully reveal his idea. Secondly, even after the product is out, since the threat of replication remains, it is necessary to strive to sustain the monopoly that being a pioneer generates. This requires 'locking in' users with the help of an appropriate user interface which they become accustomed to and are reticent to migrate away from, and locking in producers of supportive software with an appropriate 'applications programming interface'. It should be obvious that sustaining monopoly to recoup sunk costs can indeed be difficult.
 
Such strategies did help the early start-ups, resulting in the jeans-to-riches stories (Microsoft, Netscape, etc.) with which Silicon Valley abounds. But more recently it has become clear that start-ups undertake innovative activities only to create winning products that the big fish acquire. This is because of the possibility of easy replication and development of an original product, which can be done by dominant firms with deep pockets that allow them to stay in place and spend massively to win dominant market shares. In the event, the likelihood that a small start-up would be able to recoup sunk costs, clear debts and make a reasonable profit is indeed low. Selling out ensures that such sums can indeed be garnered.
 
Given this feature of the software products market, it is not surprising that small players (such as Netscape and Vermeer Technologies that delivered Frontpage) are mere transient presences in key areas even in the developed countries. To expect developing country producers to fare better is to expect far too much. The latter can merely be software suppliers or outsourcers for the dominant players.
 
What is disconcerting is that even as outsources India still remains a lower-end supplier of IT-enabled services. This involves not just body-shopping in the form of temporary export of software professionals to undertake specific jobs in large projects designed and executed in the West, but the sale of cheap skilled and not-so-skilled labour services whose output is transmitted via modern commucation technologies to sites where those services are required. That is, there are a whole host of services that can now be provided without having to move the person supplying the service to the point where the service is being delivered. Typical examples are a range of backoffice services and medical transcription.
 
The possibility of such service delivery has helped India circumvent the obstacle to service exports created by immigration laws in the developed countries. That is, a large part of software exports is not very different from the exports of nursing, carpentry, masonry and other such services, except for the fact that unlike those exports, the presence of the service provider at the point of sale is not required in the case of IT enabled servcies.
 
Seen in this light, conceptually, India's software thrust of the 1990s is not as spectacular as it appears. It is substantially a technology-aided expansion of the of the waves of migration by services providers of different descriptions: doctors, nurses, and blue-collared workers of various kinds. An expansion of that kind cannot be the basis for the redressal of international inequality that it is made out to be. But that is not all. Even in quantitative terms this development is not spectacular. The 'net foreign exchange revenue' to the country from migration of the old kind, captured by the volume of remittances into India, is in the range of $10-12 billion. The gross foreign exchange revenue from software exports is just $4 billion. The real question is, whether India can continue to exploit the possibilities of arbitrage generated by differences in software and IT-enable service costs between onsite sources and offshore centres, so as to more than triple its current gross revenues. That possibility is threatened not merely by the emergence of alternative offshore centres, but also by a reduction in the onsite-offshore cost differential. It is for this reason that industry-insiders keep emphaising the importance of “moving up the value chain” in software production. But that movement is possible as structurally constrained as the effort at increasing sophisticated manufactured exports from the developing to the developed world was.