They are still random in that the effects are unpredictible. The benefit seems to be that if you find yourself in harsh conditions, you can either die or take a chance on evolutionary change, which could be beneficial or not. If the cell is likely to die anyway, why not take a chance on change? It is not at all clear that the study you cite reports that the mutations were always beneficial.
I never said "always beneficial", I used percentage for that reason. And this conclusion comes from reading MANY papers about adaptive mutations. And again I know what you mean by "random" but this is not how "random" was/is understood. I'll try finding some article that talks about this later.
Peace be unto you