If you benchmark a product and it does 13,000 req/sec (or whatever), then answer the question of why it isn't doing 130,000 req/sec. It's a way of encouraging you to debug it further.
Got it, thanks! For me the better question is often simply "Why is it X?" Knowing why the speed is X is often of much greater utility than simply knowing that X is the speed. I like to focus on determining the current single limiting factor before considering the harder question of what would be necessary to achieve 10x.
