I worked on an idea some years ago for a couple of months until putting it up on a shelf, (beyond my capabilities) after the workable way forward was for sites themselves to identify what best labels would cover each page.

Nonetheless I slowly deduced, apart from clear spam, people would be saved a lot of time in searches if two main types of site types could easily be identified in search results, and either include or exclude these results depending on the nature of their search.

The fist being billboard or banner types, where a business had thrown up a large looking site, but really has no working data apart from address, contact, about info, quick really you knew this summary of their organisation or company.

The second are what I refer to redirection type sites, they are sites that actually don't have any / much of their own data, they're just coasting on already existing services [this might have caught people who refashion google maps with additional overlays, but so many now do not,] or an indirect way to get their parent services out thought children sites. I'm one who'd search excluding both if I'm after hard information. Generally people can use regular searches to get address and contact phone numbers for physical sales and service outlets.