I suppose there's some parallels to be drawn to software. I've often heard things like targeting 80% of memory / CPU / networking so that you have some slack fo

I suppose there's some parallels to be drawn to software. I've often heard things like targeting 80% of memory / CPU / networking so that you have some slack for unexpected events. It seems like Southwest focuses on running as close to 100% asset utilization as possible. It's not necessarily optimal when there are issues, but I find it hard to fault them for what I consider to be a pretty legitimate business choice.