At a previous employer this was a problem we identified (and larger retailer customers) had recognised, although for other reasons. We had developed a size recommendation system for them, that used real product measurements in every size and a method of obtaining your body measurements from fully clothed photos. We also offered a statistical average measurement set for those who couldn’t/wouldn’t take photos of themselves (privacy was important to us, and there was no need to undress).

We were able to give details about fit comfort across many measurements for each size, but this feature was basically unused. 99% of users used the statistical average body of themselves instead of themselves, which actually exacerbates the body type problem.

Another interesting thing about the industry and the grading process we learned; many retailers had no measurements for their own clothes except the reference size. This was much more common of higher end brands.

1 last thing; some global brands actually have the same size name on the same product represent a different size in different region (eg an SKU in size S in US may have different measurements to the same SKU in S in Asia)