As a website developer, I build a lot of websites. None of those websites are really related in an industry sense to my own business: I build websites; my clients fix automobiles, heal the sick, sell real estate, etc. I know that Google doesn't find any relevance in that respect. However, all the websites I've built are optimized for local searches using the city name(s) and metro area name in the keywords. Since my website is also optimized for local searches, will Google find relevance in the sites when relating them to my website? All the websites I build have backlinks on every page to my website (some of them deep links to articles). I know that we really have no idea what the Google algorithm is, but as someone with a programming background, I know that something like a relevance match is probably performed by parsing the text and looking for matching words, keyphrases, etc. Any thoughts on this? Should I continue to assume that I'll get at least a bit of relevance from the websites I build for local companies?
It all depends on the anchor text of your client's websites. The surrounding text and the context itself of the page will have little impact of the relevancy that the link will bring you as far as the SERP is concerned.
So if my client's websites have lots of backlinks using regional city and town names, and the anchor text for a lot of my backlinks are the same, will Google see a similarity?
As far as I know there the connection isn't made. it's a great question though, and a very interesting topic. If this were to happen, it would bring a completely new perspective to the relevancy and overall quality inbound links. One would take into consideration who is linking to us, but also, who is linking to our backlinks.
In theory, every link that doesn't come from a "bad neighborhood" (Google's words, not mine) helps a little. You won't be penalized for irrelevant links, but they don't help you much unless you have huge quantities of links.
How does Google know that they aren't at least minimally related? Hmmmmm. For example, I do web design in Boise, Idaho. One of the keywords I've started optimizing on is "Boise Web Design". I have a client that is an auto repair shop. I've optimized them for "Boise Auto Repair". I have another client that sells life insurance, and I've optimized for "Boise Life Insurance". What if the keywords were different, but used the same pattern as I illustrated above? For example one site uses the anchor text "Car Repair Service", and another uses "Car Wash", doesn't Google see some relevance if one links to the other? What's the difference between "Boise" and "Car"?
Because I don't believe that the crawling process works that way. When the bot crawls your site it follows the links within the code, it establishes a relation between the links and the content of the site itself. I don't think you directly benefit from the link chain of the site's that are linking to you, relevancy wise. Of course you benefit from the weight of the inbound it self itself, that is naturally directly related to the relevancy and authority of it's inbound links.
But the links and the content contain a lot of the same words and phrases. The words "Boise", "Idaho", etc. that I'm optimizing on are used in the pages where my IBL come from. Is Google smart enough to figure out that those terms are geographic instead of contextual?
Your question hits the core of Google contextual matching mechanisms. Finding that out is like miracle. The robot got too smart that it's probably way too hard to find out "how it works" now. By observation, it recognizes similar words or phrases, as well as singular or plural words. Relevancy is probably mainly based on synonyms, I think.