Better to be early that late - we can always slow you further or vector to lose a few miles, but it can be rather hard to make up time once it's been lost. Depending on the traffic ahead we might be able to let it run (or not).
Maestro works by calculating what everyone's untouched landing time is (based on descent profiles, winds and typical times from fix to the threshold for type and airline). That gives the landing order. It shows us the untouched time at the fix and we can accept it or adjust it (possibly shuffling the sequence) by putting in a time you tell us, a time we think you'll do (based on observation of previous traffic) or we might fiddle it a little for ease of traffic processing and overall efficiency or to fit medical traffic and such.
An acceptance rate is also set - so many per hour or so many seconds between arrivals (depending on wind affecting time spacing on final, weather, arrivals or departure mode and such). Maestro simply takes the first aircraft, adds the time spacing between subsequent arrivals and back calculates a time for you to cross the fix at.
As you can see if you miss your fix time you'll end up out of position relative to the aircraft around you in the sequence. You're 60 seconds late and the next guy is 60 seconds early and left untouched you'd be maybe 30 seconds apart at the threshold. Sometimes Maestro does a less than stellar job (actual winds different to those used for the calculations for instance) and it generates dead heats from different directions, but that's what approach gets paid the big bucks for!
You'll note we only give fix times in minutes while Maestro calculates in seconds, meaning there's already 30 seconds of play in the system so even if you're on time to the second you may get tweaked for optimal spacing on final. For better or worse it's a dynamic environment with different aircraft flying different profiles, so despite everyone's best efforts tweaking will be required. We're not doing it just for fun believe me!