Dominant designs tend to help everybody. Good ones reduce production costs, benefiting suppliers and competitors as research and development costs are reduced to near-zero. They also greatly improve the situation for complementors, who are able to produce a wide variety of easily standardized, low-cost complementary products. Consumers benefit too, as costs are driven down by fierce competition to produce the dominant design at lowest cost.
The process at its best can be seen in one of the most commonly used and least commonly thought-about products in the world: the lightbulb. There are thousands of brands of lightbulbs produced using a standard design, yet they are entirely indistinguishable. Lightbulbs from China, from Korea, Vietnam, and even America are all produced at roughly the same cost to exactly the same standards. The main variability between them, wattage, is clearly marked and well-understood. Costs are low, and innovation survives.
The dominant design in the lightbulb consists of the screwing-in part and the wattages. However, lightbulbs can be developed at all shapes and sizes, and with different color filters. Key innovations, like compact florescent bulbs and floodlights, were seamlessly integrated into the existing system. Even exotic designs, including LEDs and blacklights, were developed using the dominant design. And, of course, the number of available lamps to house them is enormous and standardized. You can even get one with a fan.
Computer operating systems, on the other hand, represent a failure of the dominant design. Because each OS behaves differently, a dominant platform like Windows forces programmers to develop their programs either exclusively for Windows or invest lots of time and effort rewriting the program. But since the platform is proprietary, it leads to a powerful monopoly on operating-system technology that creates excessive costs and relatively poor performance, since there are no direct competitors. However, I believe this situation is rare.