When measuring something like how fast a website responds. I might want to see how fast it is for 50% of people and then the slowest 25% and then the slowest 1%.
The 50% might be 120ms, but the slowest 25% is 800ms, and the slowest 1% is 2000ms.
If I graph those on a normal graph variations in the 50% line are lost because of the scale needed to show the 1%. if I graph at a log scale I can see the detail for all 3 of those lines.
Latest Answers