Chapter 89. Describing Performance Improvements
Once I saw an ad claiming that a product could improve response times by 1,000%. Sounds great. But what does that mean?
Here are two ways to express response time improvements:
- What proportion of the time did you save?
- p = (old − new)/old. It’s the improvement expressed as a proportion of the old value. You can multiply p by 100% if you want, to express it as a percentage.1
- How many times faster is it now?
- That’s n = old/new. Don’t express this number as a percentage. It’s not helpful to the recipient of the information.
Here are some examples:
| Before | After | “p% faster” | “n times faster” |
| (old) | (new) | (p = (old − new)/old × 100%) | (n = old/new) |
| 60 sec | 60 sec | 0.0% | 1 |
| 60 sec | 59 sec | 1.7% | 1.017 |
| 60 sec | 20 sec | 66.7% | 3 |
| 60 sec | 1 sec | 98.3% | 60 |
| 60 sec | 0 sec | 100.0% | +∞ |
For throughput, just use the reciprocal of the n formula. For example, if you improve throughput from 1 thing/sec to 10 things/sec, then you’ve increased throughput by new/old = 10 times. As with n values, don’t express this value as a percentage.
Note that a p value can never be bigger than 1 (or 100%) without invoking reverse time travel.2 So, when someone tells you that their thing “improves response time by 1,000%,” then you know they’re not telling you a p value. It’s probably an n value, meaning that their thing really makes your response time [merely] 10 times better than it was.
To say what sounds better, even though it’s misleading...is not your best pathway ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access