The actual physical activities the restaurant does, such as choosing a menu, cooking food, handing food to customer, playing low music, cannot be good or bad on some independent scale. What you may see as bad service, I may see as great service.
I think the only possible thing is “what does the market think.” So the entire problem is correlation.
“Bad service” is, literally by definition, the name we give after the fact to types of service that correlated with bad reviews and low revenue.
at a minimum, the health department disagrees, re cooking and delivering food.
Once we went to a restaurant with hardly any diners, not at all busy. We ordered dinner. They served half our order. We waited for the other half. After about half an hour, we asked the manager. He went to investigate, and came back with the admission that they had forgotten to submit the other half of our order to the kitchen.
That's objectively bad service. I don't see how anyone could see that as "great service".
Now, maybe it was just an innocent mistake, someone having a bad day, and maybe the manager is really apologetic and offers some compensation and I forgive them. It would rule them out of consideration for a five star review, but I wouldn't give them one star just for that. Unless, if I had gone on a review site and saw a dozen other people complaining about the same restaurant messing up their order, then I probably would have left a one star review too.
The question was - how do they know it's the bad reviews that hurt the restaurants, or if it's actually the bad service (which caused the bad reviews) that hurts the restaurants.
They observe an effect - hurt restaurant - they see bad reviews - they blame the bad reviews. But was it actually the bad service (evident from the bad reviews) which hurt the restaurant?