The retail paradigm is responsible for less than 30% of all software development. The cost of the software development is usually borne by a single manufacturer and the risk of failure is borne entirely by the manufacturer. The cost of development is distributed to customers if the product is successful on the market. Since it cannot directly distribute cost and risk until the product matures, manufacturers turn to investment markets as an external mechanism to distribute cost and risk. The need for outside capital is of long duration. Successful companies may re-invest profits in the development of subsequent software products rather than return to investment markets.
The overhead of the traditional brick-and-mortar retail sales paradigm is extremely high, with less than 10% of the money paid for the software by the end-user actually goes into product marketing, software development, and documentation. Microsoft spent only 16.8% of its 2004 revenues on R&D. The rest went into items such as the very expensive process of finding customers for Microsoft’s products: advertising, design and manufacture of an attractive package that that is discarded after the sale, payment to the retailer for shelf space, sales staff, and profit.
The retail model can only be used economically to create products for a mass market. A mass market will mask the inefficiency of the paradigm, as each customer can pay a relatively small cost compared to the cost of software development. Many software products cannot be created because they would not provide a large enough market to amortize both the cost of development and the large overheads. Moreover, many new products that could eventually build a large market will not be considered, because companies and investors cannot be convinced that such a market would develop, or the risk of failure is too high. This tends to damp innovation within this model.
The most important innovation in the last decade of global computer business, the web server and the browser, had to be developed as an open source product at a federally-funded university research laboratory. No company was convinced that it would be profitable. The only company that did invest in the project (Autodesk, Ted Nelson’s Xanadu Project) chose not to complete it due to the uncertain future revenue model.