Yet this is what we do at every election.
It is true that lies are about different things and have different impacts. People value different things. Lies are difficult to enumerate. Yet if person A says he ate 99 jellybeans when he only ate 98 and person B says he ate 99 jellybeans when he actually ate your mother, we can safely say that person A's falsehood is sufficiently different from person B's that we would rather put A in charge than B.
You can make this abstract argument, but imagine you have two politicians, call them T(rump) and H(arris). You ask advocates for T and H to make lists of the others' politician's lies ranked from most to least egregious and present them for judgement to a third party who will make an effort, showing their work, to evaluate their truth. Neither party may be fully satisfied by this process, but one is going to look ridiculous much faster than the other. When the argument is particular, not abstract, it becomes much harder to sustain.
Throwing your hands in the air and giving up because we cannot calculate a numerical measure of honesty like we can digits of pi is silly, because we still have to make decisions based on imperfect information. Claiming one must do this -- give up on evaluating honesty -- while still advocating for a particular candidate or policy is itself dishonest: you clearly do not believe your own argument but propound it in the hopes that others will and will then withdraw from the debate.