I'm not particularly fond of the entire potency per second math surrounding cooldown usage. It works perfectly fine if you assume the fight duration to be infinite, but realistically, you have a general idea of how quickly your group finishes an encounter and you can build your cooldown usage around that time frame.

For instance, let's say you're doing T10. In T10 you generally always have something to hit so you can pop cooldowns as they come up if you so desire. Let's say your group took around 390 seconds to clear in recent runs.

If you were to pop blood for blood any time you can, you'd pop it 0, 80, 160, 240 and 320 seconds into the fight (I am aware that nobody actually pops it right on pull but that's not the point here. you can change the numbers as needed). In order not to lose any blood for blood uptime, popping it 370 seconds into the fight is the latest point in time to do so. What that means for you is that there is a 50 second window (pushing the last blood for blood from 320 to 370 seconds) across the entire fight where blood for blood can be up and you can delay using it without any consequences as you don't hit the 370 second mark on your last blood for blood usage unless you go beyond 50 seconds. For this particular fight duration, you could most likely set up a perfect or near perfect sequence of GCDs for each of your blood for bloods - losing absolutely nothing in the process.

You'd probably want to assume that there is a window for the fight's duration (for instance, 380 seconds to 400 seconds if your group tends to finish at around 390 seconds) but I believe that planning out your cooldown usage based on fight duration will yield better results than assuming that each GCD you delay cooldowns by results in an inherent potency per second loss and cooldowns should always be popped within 2 GCDs.