Picture money that grows continuously at 100% a year, so that after t years one dollar has become et dollars. The natural log runs that question backwards: lnx is the number of years it takes one dollar to become x dollars. The dollar reaches about 2.72 dollars after exactly 1 year, so lne=1. At time 0 it is still 1 dollar, so ln1=0. It doubles after about 0.693 years, so ln2≈0.693.
Because ln and ex undo each other, ln(e5)=5 and eln7=7. That is the reason to use it on an equation like e0.3t=4: taking ln of both sides strips off the e and leaves 0.3t=ln4, so t=0.3ln4≈4.62. Logs also turn multiplying into adding, so ln6=ln2+ln3: growing by a factor of 2 and then by a factor of 3 takes as long as the two times added. And lnx exists only for x>0, since et is never zero or negative.
The common log, written log, looks the same on the page but uses base 10: log100=2 because 102=100, while ln100≈4.605. Read the base, not the shape of the symbol. A model built on e calls for ln, and a scale built on powers of 10 (decibels, pH, earthquake magnitudes) calls for log. Either one can solve an equation in the end, but only the matching base cancels with no leftover conversion factor.
Think of it like this. ex works like a clock that reports how much continuous growth has piled up after x units of time; ln reads the clock in reverse, reporting how long it took to reach a given size. Sizes below 1 give negative times (the past), and sizes of 0 or less are never reached, so ln is undefined there.
Core idea. lnx asks how long continuous growth takes to reach x, and it undoes ex.