The coefficient of performance of a heat pump isn't a measure of thermodynamic efficiency, because as you pointed out, a heat pump moves heat rather than creating it. That's why you get more heat per Watt-hour, because you aren't expending energy trying to make heat.
The closest analogy that I know of is in electrochemical processes where some of the energy input is allowed to take different forms. If you put in 85 units of electrical energy, 15 units of waste heat from a different process, and get 90 units worth of product, your process is 90% thermodynamically efficient, but has better than 100% electrical efficiency. In the case of a heat pump, measuring units of heat per Watt-hour is a bit like this in that some of the input energy (for the heat generation) has already been provided by a different process.
The closest analogy that I know of is in electrochemical processes where some of the energy input is allowed to take different forms. If you put in 85 units of electrical energy, 15 units of waste heat from a different process, and get 90 units worth of product, your process is 90% thermodynamically efficient, but has better than 100% electrical efficiency. In the case of a heat pump, measuring units of heat per Watt-hour is a bit like this in that some of the input energy (for the heat generation) has already been provided by a different process.