The weird thing about these is that they blame Google's search results on spam. I work in SEO and I can tell you that they are much better at ignoring spam than they were in 2010, where a lot of these people quoted still have their heads at regarding SEO.
What's been going on at Google is reliance on neural nets to take care of various ranking algorithm tasks. We want better keyword matching to generate results, but Google is developing ways to match query vectors to document vectors using stuff like BERT. Google is looking at the knowledge graph of entities that emerges out of the content we write and is trying to figure out which relationships between entities are important to a query and which result set has the best coverage and diversity. This incentivizes publishers to write a lot of text that covers multiple related topics and bury the point inside of it.
The other major shift in Google is how they consider links. PageRank is still around in some form, but there could be other link-based algorithms that serve similar purposes. The last few years of core algorithm updates put a lot of importance on receiving links from news websites for any keyword with commercial intent. If you want to rank, go hard on public relations.
The result is a real loss of accuracy and a lot more false positives that are semi-related to the query.
>This incentivizes publishers to write a lot of text that covers multiple related topics
Is it accurate to call organisations that write text according to google's incentives 'publishers' or are they merely spammers trying to maximise their pageviews and conversions?
Yup. IMHO spam has become so good at mimicking genuine content, it's hard to recognize even for a human curator. There's so many websites in the top google results that I'm sure are entirely AI generated, which exist for the sole purpose to propagate affiliate links and ads.
Yes. It's like the results when people realized you could have a classifier trained to match a person's face, reversed to generate a new face based on the classifier. There are a few extra steps, but the web is just recipe sites and product reviews that look like what the google ranking algorithms idealized site looks like.
What's been going on at Google is reliance on neural nets to take care of various ranking algorithm tasks. We want better keyword matching to generate results, but Google is developing ways to match query vectors to document vectors using stuff like BERT. Google is looking at the knowledge graph of entities that emerges out of the content we write and is trying to figure out which relationships between entities are important to a query and which result set has the best coverage and diversity. This incentivizes publishers to write a lot of text that covers multiple related topics and bury the point inside of it.
The other major shift in Google is how they consider links. PageRank is still around in some form, but there could be other link-based algorithms that serve similar purposes. The last few years of core algorithm updates put a lot of importance on receiving links from news websites for any keyword with commercial intent. If you want to rank, go hard on public relations.
The result is a real loss of accuracy and a lot more false positives that are semi-related to the query.