100% utilisation is not (generally) a sign of efficiency.
[You can take this as a literal rant about bus service, or as a parable on effective assignment of computing or business resources, as you wish.]
Once again this morning, the bus to work was full to capacity. Standing room only, everyone stuffed uncomfortably close together.
That means we're getting maximum usage out of the buses, yes?
Well, maybe.
How about I put it another way...
Every so often, I get to the "Park and Ride" and cannot park my car, because the lot is full. So I drive away, and spend at least as long in my car as I would have on the bus - but with many, many times more pollution, and an inability to read my email or work on my blog.
Clearly I'm not the only one driving away, so clearly, 100% utilisation of the Park & Ride means that drivers are hitting the roads in their cars for a long, slow, stop-and-go drive to their work parking lot, which also fills up more than it should.
The same is true of the bus service itself. There are times when I think to myself "do I want to stand crammed on a bus unable to move my arms enough to take off my coat though I'm sweating like crazy? Or do I want to sit in the air-conditioned comfort of my car, where I can listen to the radio and drink my tea without being jostled by the next guy along?"
So, despite myself, I will occasionally drive, even though I would rather ride the bus.
The same is going to be true of any service - if it is packed to the gills providing service, then it's not only going to be killing its own usefulness when there's a "blip" or spike in demand, it's also going to be killing its usefulness on a regular basis with people who try it out, and realise it's just not quite as comfortable a service as they would like.
You also can't measure how much service you aren't providing - did only one person miss your bus because it was full, or did twenty? Maybe there's a hundred more riders - three bus loads - that you could serve every day if the buses came more frequently, and people knew there was a good chance that they could sit down in some comfort, or if the parking lot at the park and ride was larger.
You can measure the number of people who were standing outside the door and had to be turned away, but you cannot measure the number of people that they spoke to and dissuaded from riding your bus.