Today I started to think how do computer-related products (CoRP) actually fare in the markets? Why does one operating system grab market share from another? Why there's a fanatic group of Mac OS X groupies? Why do people hate Windows, and why do some think it is the only possible choice?
I've got some ad-hoc assumptions about the markets in which computer-related products live. First of all;
- there's a limited number of buyers
- buyers can buy a set of 1-N different products
- not all products are compatible with each other
- companies engage in a schizoid competition: they are co-operative, but also cut-throat as soon as it becomes possible to eliminate competition
- we can model a product's lifecycle as linear; it improves in some way by each turn of the game (reality is certainly not this simple)
Problems in our model (to be solved)
- how is the mapping done from all the possible parameters describing a company's character and its products, into a curve in xy-plane?
- what is the market impetus, ie. approximately the "force that in good winds makes every player fly", but when there's a hardship, also makes some products and possibly companies fall down
- and perhaps one of the most important question: how can we do early prediction of the winners? Everyone wants to invest in as early stage as possible to the winner, but usually in the beginning every player looks approximately of equal strength
Let's get some real-world examples into this model. I'm personally very interested in the background factors affecting the selection and suitability of operating systems.
Windows is the mainstream OS, while Linux, Mac OS and some others have also their niches. UNIX is the historically preferred operating system in industrial settings. Each of these have their own characteristics, and indeed these characteristics are not so easily quantifiable. We can take price, and some other "easy" ones, but then the real questions which computer administrators and end users are interested in, are not so easy to evaluate.
There are several return-on-investment (ROI) calculations of different operating system choices. These have been a good guiding rule of thumb to those that are implementing OS use as part of their daily work.
Also the weight given to a particular characteristic can change in time. For example, when there were only perhaps one or two servers in a company, back in 1980s and 1990s, it didn't matter so much how easy it was to install or what kind of administrative work the operating system required afterwards. But when companies started deploying more servers per capita, these ease-of-use and deployment questions began to be much more acute. In fact, companies could react with basically three alternatives: put more bang into one server (meaning more roles or features), ease the installation of a server (to enable an admin to install more servers per unit time), or develop or buy third party applications that amend the original flaws in the server operating system.