Use the Brambleberry fragrance calculator! It gives you a good place to start and a choice of low / medium / strong.
https://www.brambleberry.com/pages/Fragrance-Calculator.aspx
I agree that it can give someone a place to start (like medium), but in my opinion its use should come with a caveat. Here is why:
Nearly all the EOs with a few exceptions show the exact same amounts for light (.44 oz ppo), medium (.68 oz ppo), and strong (.80 oz ppo). That includes such different EOs like lemongrass, lavender, peppermint, patchouli, cedarwood, eucalyptus, juniper berry, etc. and yes... black pepper! The ones that differ are mainly the citrus EOs which call for nearly half the amount. What's wrong with this picture?
The components that give EOs their characteristics are volatile, meaning they evaporate into the air at room temperatures, otherwise we'd never smell them. Those molecules actually travel into the air to our noses. Once evaporated, they're gone. So the strongest smelling EO soap may contain elements that evaporate very quickly and nearly overwhelm us at first, but later leaving you with a practically scentless soap after being on a curing rack for a couple months.
Also the calculator doesn't take into account individual differences in actual scent, and how they're received by average people at various strengths. For example, a "strong" lavender @ .8 oz ppo might be great in a bar after a 2 month cure. The same strong amount of patchouli or eucalyptus might be sickening to some. Or the opposite may be true. Some of these scents are easier to take at higher levels than others.
Not the mention the whole light, medium, and strong levels are rather subjective, and cover as big a range as I mentioned (100%)! That's why I think it best to post a question about the specific EO and let a dozen people chime in with the amounts they use, along with comments on how the bar smells once completely cured and sitting around for awhile. Of course do a forum thread search first in case such an EO has already been talked about at length.