Here in the Southern parts it works fine, but I do switch to oil, below approx. -17 (when COP < 1.5). Still, one of the most important things overlooked by those installing air-to-water units is the total surface area of the distribution (radiator size, or in-floor heating vs. not having one). It completely changes the game. If you need to run 60degC water into the system at mere -10degC, you can't win. Even ground source would struggle to give good savings but it at least somehow works. But the whole game changes if you run 30degC water in massive radiators or in-floor heating. I invested 3000€ in the heat pump and another 600-700€ into upgrading radiators and it's working well and based some modeling and approximate measurements I'm expecting SCOP significantly exceeding 2. The radiator upgrade pays for itself in a year or two. Such "small" details are hugely important.
Yes, I upgraded radiators as well when switching from oil to ground heat pump (the bigger the better). In-floor heating would be best, but this is an old house, so not that easy to convert. Now it gets to 55-60 ℃ at -25 ℃, so kind of fine. Those temps are rare, but happens about once per winter.
Some statistics I've checked/calculated from the heat pump:
COP 4.4 at 0 ℃. Heat output 2.4 kW, electrical power 0.54 kW.
COP 3.9 at -15 ℃. Heat output 4.5 kW, electrical power 1.16 kW.
(this must be momentary numbers, my notes are a bit messy)
With a newer house with better insulation and only in-floor heating, those numbers would be better.
Edit. Looked up my notes and it looks like SCOP improved from 3.2 to 3.7 due to the larger radiators. In 2021 SCOP was 3.7. These numbers are quite reliable, because I've a separate energy meter at the pump and the pump itself has its own flow sensor, calculating total heat output.