Why AI Content Rises, and then Falls?

? Subscribe here to weekly Marketing Shots and follow me for more Nick Malekos.

? What you will read:

?? Why AI sites grow

?? Why they fall

?? The problem with Site-Wide Quality

We have all seen the case studies of high SEO growth by companies starting producing hundreds of pages of SEO content.

Below is https://history-computer.com/, which started producing hundreds of AI-generated articles in 2022, then thousands on 2023.

Why do all these sites grow so fast, then crash so hard?

Growth Factors

First-chance indexing: Every new page being indexed, is initially being viewed more favourably and is given a chance to be tested in SERPS and collect clickstream data. That means, many new pages equals more visibility as soon as they are published.

Publishing velocity & topical authority: Publishing many pages on the same topic creates authority on the topic, what we call topical authority, pushing all the other pages up as it goes. High velocity of pages & articles helps these sites grow.

PBNs and other shady links: In some cases, these sites use Private Blog Networks to boost the domain or the pages. PBNs are often hidden (blocked) from SEO tools like Ahrefs or Semrush, so they won't be visible in your reports, but are visible to Google, pushing the pages up on SERPs even when they have bad content.

P.S.: The problem with viewing Ahrefs data for these sites is that Ahrefs does not update all the keywords at the same time. We actually see a slower growth, steadier peak, and gradual fall. It's possible that growth & fall is much faster and hard.

Fall Factors

Algorithm updates: Most of these sites got hit by the newest algorithm updates (Core+HCU) by Google and you can see their drop begins late on Summer 2023. However, they share so many common characteristics

Page quality, clickstream, & de-indexing: Another factor is page quality. If a page is terribly written by AI, then users do not stay there, and possibly Google can understand bad content with or without AI assistance. As those pages collect clickstream data (people not clicking or leaving after visiting), those pages start to get de-indexed, lose rankings, and keywords, leading to a slow decline.

Penalties: In some cases, these pages may experience a manual penalty such as for using a PBN that was discovered or another shady tactic.

Quality score & formatting: I keep seeing the same "blocky" formatting in articles written with AI, huge paragraphs, no bolding, no internal/external links, no subheadings (after H2), or bullets & visuals. These shout low-quality AI content.

Cost of maintenance: Content in competitive niches requires updates often. You can't do that if you have a team of 2 engineers and thousands of pages ... In most cases, you can't update or keep "fresh" and relevant the pages for long, until maintenance dept crepts in and pushes you down.

Site-wide quality

I wanted to focus a bit more on Site-wide quality.

Sites that were previously ranking well see drops due to AI content because of lower site-wide quality.

Google has many ranking factors. They have ranking factors for individual pages, but also ranking factors for whole websites.

When you are publishing decent or good pages for a while, you create a high-quality expectation, and your new pages get the same treatment.

But, when those new pages keep being low-quality, you reduce the average quality score of the whole site.

In the case of AI, thousands of bad pages and a few hundred of good ones, creates a precedence of a very low quality score which drags your whole website down.

Final Words

AI is not a bad actor here.

People who want to get rich quick without doing the actual work are the bad actors here.

SEO content needs work, and AI can assist in producing content faster. But, it requires monitoring, editing, and proof-checking, not to mention formatting, visuals, and some good old fashion personal experience!