I have a feeling that the manufacturers have, over the years, managed to learn that after you have designed a motor, you still have to test it to see if what has been designed is, in fact, what is being produced. There are plenty more variables involved over and above those mentioned - brush quality and air gap being another couple. The empirical method, while messy, is far more reliable than using a manufacturers details that might well reflect a perfectly bedded in motor with an unaged magnet. Better to guess, get a bit of surplus "go", and run with the wick turned down a bit than select a motor that might, under perfect circumstances, only just do the job.