Most consumer input devices are based on boiled down micro-controllers that only handle the bare minimum of tasks, so an EEPROM does become another cost that could add up when you mass produce.
However, the few cents it takes to use a slightly more capable controller is of negligible consequence to the end-user price, the real cost is that you need to spend a sizable amount on software development; i.e. custom communication protocol for programming the device and a user interface to do the programming, not to mention the underlying driver tweaking, testing, etc...just making it all work reliably and consistently on different versions of windows or other OSes could make up 25% or more of your cost to develop the device, and then you still got to have a market to sell it into...and as Rajagra mentioned, you need additional support, particularly if you cut corners in the software development area to begin with.
People that would really get the most out of such a device have the money to pay for the ones that do exist, i.e Retail POS, or more likely, they skip the programmable part and just get something made for them, i.e. the
Bloomberg keyboard