>If I just measure according to the rotations of the spool being spooled onto, this wouldn't at all consider the spool getting thicker with the thread already on it.

Is the thread always the same thickness?

Measure the thread some other way, wind that thread onto the spool noting how many turns it turned, and use that from then on.

Run some tests with thread of known length -> "X turns does what I want it to do" -> Do it X turns from then on.