The public needs to truly learn how to wear masks, which masks to wear, etc. I bet you the efficient of, say, cloth masks is super low due to the fact that people wear them incorrectly, don't wash them properly, keep touching them, etc.
Preventing these imperfections is extremely important in a very high infectivity environment like a hospital's Covid ward, but they don't really change the masks' impact on the epidemiologics of supermarket trips.
A protection level that only shifts the average time to infection of a Covid ward's staff from Monday to Thursday would be terrible, but the same level of protection added to otherwise unprotected supermarket goers can easily reduce the supermarkets' contribution to the R value from heavy to insignificant.
Yes, of you do not impeccably follow the procedures required by medical staff in high infectivity environments the small remaining risk of defense penetration will rise by a high factor, no, the remaining protection will still be much better than no mask.