The ketone washes will separate out in 3 minutes if you double the amount of DCM. Small amounts of DCM will recover all of the ketone, but they make the washing a bear. Another thing that will help is if you do a plain water wash for your first wash, followed by NaOH washes.
This reaction scales up nicely IMHO. you can get ~50g ketone back from 97.2g of isosafrole. That's only around 40% yeild, but I think the yeilds could be higher. at a 6x scaleup, I somehow always managed to end up with a >20g forerun of isosafrole. I assumed it was isosafrole because it came over in the right temperature range, and when fed back to another Oxone/hydrolysis reaction would produce more ketone. One theory I had was that it might not be a problem involving the amount of oxidant but instead a problem with methanol/water ratios. I've wondered if adding a bit of acetone would help the mixture become more biphasic or not, but I no longer have the ability to experiment and find out. If anyone doesn't mind playing around with your solvent ratios a bit, please do and report back with the results.
The precipitate salts are easy to deal with. Just decant off your liquid, extract the remaining amount with 3 methanol pulls, add these to the mother liquor, and then add a little more than the required amount of water. Another way to do it, if you're using a smaller separatory funnel and have to do multiple extractions anyway, is to decant off the mother liquor, extract this, then go back and do you methanol pulls of the precipitate salts. Take that methanol and add 2-3x its volume of water and extract this with DCM separately. Doing it this way prevents any salts from falling out of solution and clogging up your sep funnel.
All paths are the same: they lead nowhere