I think this is exactly what I investigated in my post here, just from a different viewpoint. And yes, this can definitely be better, depending on pump power requirements and actual heat loss. Your approach of adapting flow rate to the actual heating load is definitely optimal - basically if you enter different load wattages into the spreadsheet in the linked post, you’ll get your optimal flow rate for all scenarios.
I am running my oversized 10 kW pump at minimum flow rate (1000 //h) for precisely that reason: I save 50W in pump power and didn’t even have to raise flow temperature because the required increase was so small it’s negligible.
I should try to also rewrite this from the perspective of “Optimizing dT for ideal COP”…