The Concepts of Increasing and Diminishing Returns
Sidebar to Jakob Nielsen 's column on increasing returns for websites .
Much of human economic activity suffers from diminishing returns . For example, in farming, the farmer will first farm the most fertile land with the most valuable crops. To expand the farm's business, the farmer will have to cultivate progressively less fertile land and will have to grow less valuable crops (once the demand for the most valuable crop has been met). In general, the bigger a business gets, the less optimal its last venture. Assuming that a manager is good at picking the most promising business opportunities to do first, it would seem that diminishing returns are a fundamental law of doing business: the first few deals skim the cream and subsequent ones have progressively less value.
Despite the general prevalence of diminishing returns, some industries behave differently. In particular, the software industry seems to work under a principle of increasing returns where getting bigger results in better deals than if you stay small. Software manufacturing (floppies or CDs) is very cheap relative to the R&D cost of developing the code in the first place, so only by getting volume sales do you make real money. Furthermore, customers find extra value from buying software from a bigger vendor: because more other people use the software, it is easier to exchange files in its data format, it is easier to hire staff that is trained in the use of the software, and it is even easier to find books that explain how to use the software. In other words, the more other people buy a software product, the more likely you are to buy it yourself.
Simply put, if you double the size of a business, does it become more or less than twice as valuable? In traditional industries, diminishing returns set in, so getting 100% bigger may only generate, say, 90% more value. In software and other industries governed by increasing returns, getting 100% bigger may generate, say, 150% more value. Thus, the question is not whether bigger is better (it almost always is), but how much better it is to be big.