Not OP, but let's say you want to find out the protein content of brussels sprouts.
I would type `protein content brussels sprouts` (without quotes) because I fully understand that the information I'm seeking might be in some tabular form, or phrased in a way I don't anticipate.
Most non-technical people however would type in `What is the protein content of brussels sprouts?` literally.
This leads content creators who see these queries in "keyword analysis tools" to dump SEO-optimized crap into millions of blog posts, with completely irrelevant word soups with countless varations of the question, and the actual information buried deep within that gibberish essay, unreadable by humans, only optimized to drive ad traffic.
Google's optimization for the non-technical use case has lowered the overall search result quality immensely.
There was a time when Google didn't simply ignore some of your search words, or when control characters like +, -, and "" were actually respected (~pre 2010), and the introduction of verbatim mode didn't change much IMHO.
Interesting. Why do you think that content creators wouldn't see `protein content brussels sprouts` in "keyword analysis tools"? Why do you think they wouldn't create millions of blog posts containing words `protein content brussels sprouts` or countless variations thereof?
Because Google optimizes for "quality content" since at least the Penguin update. They're using NLP tools to assess the writing quality (similar to algorithms telling you at what school grade level your writing is).
This was good, because it cured all the copy&pastable 2000s era "tag cloud" sites which simply dumped tons of search keywords all over the place.
Ideally, it lead to a stronger emphasis on high-quality human-written content, but it turns out that this algorithm, again, is easily fooled by feeding it "SEO essays" that looks like prose, but is irrelevant text gibberish, but written coherently.
That lead content creators to expand data that would ideally be presented in tabular form on one page, to multi-page "SEO prose" that looks like it's written for humans, but is completely undigestible.
That, along with Google's auto-suggestion feature that finish your sentences after you type in some words, especially on mobile, lead to the impression that people actually like to search in full sentences.
One can easily make them minority by creating a script which issues keyword-only searches to Google and let it work 24/7/365 from a few hundred machines.
I would type `protein content brussels sprouts` (without quotes) because I fully understand that the information I'm seeking might be in some tabular form, or phrased in a way I don't anticipate.
Most non-technical people however would type in `What is the protein content of brussels sprouts?` literally.
This leads content creators who see these queries in "keyword analysis tools" to dump SEO-optimized crap into millions of blog posts, with completely irrelevant word soups with countless varations of the question, and the actual information buried deep within that gibberish essay, unreadable by humans, only optimized to drive ad traffic.
Google's optimization for the non-technical use case has lowered the overall search result quality immensely.
There was a time when Google didn't simply ignore some of your search words, or when control characters like +, -, and "" were actually respected (~pre 2010), and the introduction of verbatim mode didn't change much IMHO.