[iansane]

there are situations where percentage is misleading.

10% of people start a startup today, 5% 5 years ago.

20% of young people start today , 30% 5 years ago.

chance of a young person starting today? 2%, 5 years ago? 1.5% . therefore young people are more likely to start a startup today than 5 years ago

[samastur]

To clarify I think what you meant in your example is that 20% of those who start today are young while there were 30% of the same cohort 5 years ago.