In The birth & death of search engine optimization, Xe suggests
Here’s a fun experiment to try. Take an open source project such as yt-dlp and try to find it from a very generic term like “youtube downloader”. You won’t be able to find it because of all of the content farms that try to rank at the top for that term. Even though yt-dlp is probably actually what you want for a tool to download video from YouTube.
More generally, most tech folks I’m connected to seem to think that Google search results are significantly worse than they were ten years ago (Mastodon poll, Twitter poll, Threads poll). However, there’s a sizable group of vocal folks who claim that search results are still great. E.g., a bluesky thought leader who gets high engagement says:
i think the rending of garments about how even google search is terrible now is pretty overblown1
I suspect what’s going on here is that some people have gotten so used working around bad software that they don’t even know they’re doing it, reflexively doing the modern equivalent of hitting ctrl+s all the time in editors, or ctrl+a; ctrl+c when composing anything in a text box. Every adept user of the modern web has a bag of tricks they use to get decent results from queries. From having watched quite a few users interact with computers, that doesn’t appear to be normal, even among people who are quite competent in various technical fields, e.g., mechanical engineering2. However, it could be that people who are complaining about bad search result quality are just hopping on the “everything sucks” bandwagon and making totally unsubstantiated comments about search quality.
Since it’s fairly easy to try out straightforward, naive, queries, let’s try some queries. We’ll look at three kinds of queries with five search engines plus ChatGPT and we’ll turn off our ad blocker to get the non-expert browsing experience. I once had a computer get owned from browsing to a website with a shady ad, so I hope that doesn’t happen here (in that case, I was lucky that I could tell that it happened because the malware was doing so much stuff to my computer that it was impossible to not notice).
One kind of query is a selected set of representative queries a friend of mine used to set up her new computer. My friend is a highly competent engineer outside of tech and wanted help learning “how to use computers”, so I watched her try to set up a computer and pointed out holes in her mental model of how to interact with websites and software.
The second kind of query is queries for the kinds of things I wanted to know in high school where I couldn’t find the answer because everyone I asked (teachers, etc.) gave me obviously incorrect answers and I didn’t know how to find the right answer. I was able to get the right answer from various textbooks once I got to college and had access to university libraries, but the questions are simple enough that there’s no particular reason a high school student shouldn’t be able to understand the answers; it’s just an issue of finding the answer, so we’ll take a look at how easy these answers are to find. The third kind of query is a local query for information I happened to want to get as I was writing this post.
In grading the queries, there’s going to be some subjectivity here because, for example, it’s not objectively clear if it’s better to have moderately relevant results with no scams or very relevant results mixed interspersed with scams that try to install badware or trick you into giving up your credit card info to pay for something you shouldn’t pay for. For the purposes of this post, I’m considering scams to be fairly bad, so in that specific example, I’d rate the moderately relevant results above the very relevant results that have scams mixed in. As with my other posts that have some kind of subjective ranking, there’s both a short summary as well as a detailed description of results, so you can rank services yourself, if you like.