Apparently the 'industry preferred' method (quoting the docs available in my aisle at work) is to allocate the time crossing the feeder fix to the pilot, with 250kts after...in practise though, If running a sequence I'd want more control over it, so i issue a descent speed, give an indication of the feeder fix time we're after, and finesse through vectors if needed. That way I avoid having a leading aircraft reducing to 230kts to meet the time at the feeder fix, only to have number 2 in the sequence at 280kts yet still meeting their feeder fix time.