Ah, but, you see, the number of pleased audience members required for "good" fanwork is absolute, not proportional, and therefore knowable! And while it is possible to survey the satisfaction given by feedback, the survey conditions will be such that response bias is inevitable and impossible to correct for.

Also I wish to point out that statements such as "authors are very easily pleased", while they contain semantic meaning and may even be empirically provable, have very little evidentiary value, as nobody who finds feedback impossible to write is physiologically capable of integrating that statement in a functional manner. :P

As for your final parenthetical: it is absolutely correct, but is unrelated to your argument, as utility and difficulty are independent variables: meanwhile, creating bad fanwork has the primary negative effect of only reducing potential audience size for one's own work (already demonstrated to be bad and therefore of negligible value) while leaving bad feedback has a nonzero chance of reducing the amount of good work created by the other fan going forward, which gives it a much more substantial risk factor associated with the increased difficulty.