Working on many robots over the last 15 years and specifically home robots the last 4; brainstorming new ideas I developed this infographic about the home robot design space. There are some sizes of robot that that are advantageous to build and some sizes you just don’t want to build.
The key insight is that once you’ve built a robot that’s tall enough not to fit under furniture anymore (i.e. big enough to navigate through the environment in a way similar to humans), there aren’t any additional big gains in capability until the robot is tall enough to see over tables, reach door knobs etc. but it get’s difficult to keep the weight (and thus the cost) down as you get bigger. So why would you make the robot any taller than you had to? In my experience, the total cost of a robot is at least proportional to weight and if you include logistics costs it’s much worse than that. As the robot get’s heavier, it needs bigger motors to move it, which makes it heavier, and bigger batteries, which makes it heaver, and a stronger frame, which makes it heavier, and now it’s dangerous so you have to add more sensors to make sure it never hits anything, which draw more power and need bigger batteries and so on. If you want to be able to make something that’s mass producible and can be sold at a reasonable price, stay as small as you can while still being able to accomplish the product goals.
Not overpromising
As an aside, another big issue with robotics is over promising and underdelivering. If a robot is more capable than people expect it to be they are delighted but if it is any less, then it is perceived as stupid and there are no second chances. The funny thing is that people seem to expect things to be about as competent as animals they are familiar with that are a given size. If your robot the size of a hamster then the expectation is that it should be about as intelligent as a hamster. But if it’s the size of a human then you’ve just set an expectation that will be very hard to meet.