On the other hand, you can say that you want the average latency per request to be under 100 milliseconds, and setting such a goal could in turn motivate you to write your frontend with low-latency behaviors of various kinds or to buy certain kinds of low-latency equipment.
