I posted this in performance as well. I had an idea, ran some tests, and cut the time in half. I imported the same 500 items and updated the same field with one change each time. My theory was that if I created a custom form that had sub lists and unnecessary tabs turned off it would speed things up. The results were 13min default form and script on, 11min smaller form and script on, 6min default form and script off, 3min smaller form and script off. The script was a user event that triggered on all three events, checked if it was in the UI, and immediately returned. So the undocumented optimization is create a form that isn’t stored or preferred, with the smallest number of fields you can get away with and you could cut an already optimized import in half.