Exactly, that "all or nothing" is the main problem in in how most of the people think about the whole topic.
I've read one work that gave the very simple estimates: if the infected person wears the surgical mask it will emit 50% less in the air, and if the non-infected person wears the surgical mask that person will get 60% of what's in the air. But combined, that means that the non-infected person will get around only 30% of what it would get when nobody would wear a mask. Which seems to be way too little for personal "full" protection, but looking more globally:
If the R factor (reproduction number) is 2.7 with no measures, and if everybody wears a mask and other limitations are respected (physical distancing, limiting the number or people in the closed spaces, ventilation, etc.), all together it can result in the R factor being reduced to less than 1: With the known no-measures factor 2.7 we have exponential growth. With the factor under 1, we have a steady decay in the number of infected people.
There we also "just" have to reduce that factor to 30% of the no-measures one. It's not so easy as some people spread to many, and a lot already don't pass the virus to others. That's why groups of people near each other change everything too much.
But just by maintaining the factor under 1 (the lower the better) we guarantee that the relative number of infected around us will be always be smaller and smaller, and that's can be more than enough for the population as the whole.
That's why wearing the masks and any additional good measures together do help, even if every measure, evaluated alone, for only one person, "doesn't protect" enough.
So it is not just either 0 or 100%, but everything in between: not everybody has to use the same protection that the medical workers have to to improve the chances for everybody. But everybody should do it. One person who's actually infected and not wearing its mask changes the number of viruses passed to anybody in its vicinity from 30% to 60%, when all the rest wear the masks, or to 0 to somebody who's also not wearing the mask. From the starting 50% and 60% of masks "working" we get all the other outcomes too.
And the same is valid for testing. Rapid wide scale cheap testing to identify who should enter the building doesn't have to be as good as PCR tests to still give us immense benefits, if we have it and use it. We just have to be clear that we want it, and that we know now it's not about "either 0 or 100%."
Also, the pictures (and the video) explaining much more than many words:
"Face coverings and mask to minimise droplet dispersion and aerosolisation: a video case study"
https://thorax.bmj.com/content/early/2020/07/24/thoraxjnl-20...