Start with a log you can do in your head: log28=3, because three doublings turn 1 into 8. Now ask the same question in base e: ln8≈2.079 and ln2≈0.693, and 2.079÷0.693 comes out to 3. That is no accident. Since 8=23, the power rule gives ln8=3ln2, so dividing by ln2 recovers the 3. The argument works for any number: if logbx=y, then x=by, so lnx=ylnb and y=lnblnx.
It behaves like a unit conversion. Logs in different bases all count how many times you multiply, but with steps of different sizes. One step of ×2 measures 0.693 in natural-log units, so to count how many ×2 steps fit into ln7≈1.946, divide: 1.946÷0.693≈2.81. This is also why every log function is a constant multiple of every other: log2x≈1.443lnx for every positive x.
The formula resembles the quotient rule, but the two say different things. lnblnx is a quotient of two logs and changes the base; lnbx=lnx−lnb is the log of a quotient and keeps the base. Numbers show the gap: ln2ln8=3, while ln28=ln4≈1.386. If a problem stays inside one base, use the log properties; use change of base only when the base itself has to change.
Think of it like this. Converting a log is like converting a trip from miles to kilometers: the trip stays the same, only the unit changes, so you divide by the size of one unit. Here the unit is one factor of b, which measures lnb on the natural-log ruler. The analogy is exact, since all logs are constant multiples of each other.
Core idea. logbx=logablogax rewrites a log in whatever base you can actually compute.