The page loads in 80 ms locally. In prod, on the real database, it takes 9 seconds and freezes the tab. The code did not change. The data did: 60 test rows locally, 14,000 in production.
The culprit was a ten-line loop, perfectly readable, hunting duplicates in a list. It had no bug. It just had the wrong shape. And understanding that shape is all Big O is about. No math, promise: just enough to read your own code and know which part will explode as the data grows.
Big O measures a shape, not a speed
First intuition to break: O(...) does not tell you "this takes 3 milliseconds". It tells you how the time reacts when you double the data. It is a curve shape, not a stopwatch. Three families cover the overwhelming majority of the job:
O(1) — constant. Data size changes nothing. Reading array[5000] is as fast as array[5].
