AI Article Penalty Speed Run Case study 4000 posts

Jaordas

Power Member
Joined
Oct 28, 2019
Messages
770
Reaction score
1,879
So, I did a thing, this is completely stupid and by no means do I encourage anyone to do this, much less on a site that actually was worth something. Im posting this purely for the few people who are going to find this interesting, I have no goals for this site, I have no real intention of making any more money with this site, I am not going to make this into some sort of formal looking post, im just going to run my mouth on what stupid things I'm doing to this site and we will all see how it turns out if I don't forget to give updates.

tldr for what this post is about before I go into it --- I Posted over 4000 AI written articles on a legitimate website I owned just to see what would happen........... I posted it all in a single week.

So, time to run my mind on some things id like to talk about with what is going on with this site, its history etc.

I started blogging etc, probably 3-4 years ago, in many ways I'm a baby in this but frankly white hat SEO has been so easy for me that this website is literally a throw away, it was my first ever website and I feel some sort of attachment to it so I never sold it, I never black hat anything on it in any way before now.
The website is in a niche with heavy seasonal, that's why you will see massive dips and jumps in traffic, meanwhile I probably hadn't published a single thing on it in 2 years before now.
All this website became was that annoying paycheck I get in my bank thinking it was from my actual money site just to be disappointed.

This website, while completely human-written before (100% by me, never hired writers because back when I started blogging I couldn't even afford that) is probably undergoing some issues, it was made by me before I knew the things I know now, and last year a google update wiped out every single featured snippet it has. For some reason it maintains rankings but gets no featured snippets, id be lying to you if I told you I tried to fix it though, I barely have time for this website.
I will probably pull the website from all ad related services since I don't want to get some sort of issue for content quality etc as the site uses the same services as my actual websites. So I likely wont have any Ezoic screenshots in updates.



How did I end up doing something like this? well, we all heard about chat gpt, AI etc, I was blown away when I asked Chat gpt for an article, it was written so well, you literally pay more for a human to write worse.
I keep seeing some of my friends telling me their AI content is ranking, and others saying it isn't. One friend had an AI site get hit, then he updated the best articles with human writing and it recovered, which is strange as 90% the site is still AI. I know so much SEOs who hit one problem and a website is dead for a year plus, the dude recovered his in 2 months.
This made me wonder if google is only messing with the AI content on a page to page basis, and you could actually just pump content out without the whole site drowning because the other articles were AI.

Another dude I know, just deleted a lot of his AI articles and his site recovered, these little things got me thinking, maybe I could try something, and see what happens on a little website I own, because people arent doing the kinda stress tests I want to see.
So I have this site, which google is already messing around with and I have no motivation to write the content for, but I DO know how to get all the keywords in the niche and have software to combine keywords that have the same search results (meaning "how to turn on a lamp" and "how do you turn on a lamp" would be combined into the most traffic one and I could avoid duplicate content and keyword cannab).

So, I start off by getting every single keyword in the sites entire niche............ I'm not joking, every last one, over 10000 keywords, is compressed down into 3900 ish posts after keywords are combined.
I was going to start off soft by doing 100 posts in a week, but I just felt like that was moving too slow, by the time I ever see results interesting enough to me, google might actually combat AI. Since I have bad luck with some black hat, like google only realizes they need to stop it the moment I start trying it on websites myself lol.
What is so interesting is people are always telling me about crawl budget, always telling me their pages don't get indexed, frankly in all my 4 short years I've never had this problem (but this is the first time I was doing volume of posts this high).

In 4 hours google crawled and indexed all 100............. I didn't believe it because based on what other people tell me, google has limits, but I gues my limit was higher???? that site is literally my weakest site so I have no idea why.
I was expecting to watch them get indexed over days of crawling. I published 100 more trying to eat up the "crawl budget" people tell me about.
All posts indexed in 3 hours...................... ranking anywhere from 5-30 for the keywords......... fully AI articles.

so I ended up publishing 1400 articles that first day, just trying to see if google would hit a limit or if my traffic would crash.

google just indexed every single post before the day was over, I even manually checked like 20 and they were all ranking (given not on page 1 but pages can take months to rank).

in 3 days I published an article for every single keyword in my niche, about 5 % of the post are already #1 for the keyword they were published for (low competition).

I am not an expert on crawl budget or things like that but its so funny to me that thousands of posts from this site were indexed and ranking in less than 24 hours.

Now just to look at the website a bit and stop rambling.
1677022976019.png

Here is just some basic exoic for it over time, its only made me about 16 grand since I made it, all the traffic and rev is 100% from posts written by me, and frankly summer and winter hate my posts.
Anything after 2021 has been the website having its content get old and lose rankings.
All those massive drops are just seasonal, except for the sep 22 drop, that was when the website lost all featured snippets across every single keyword.
The nice has low epmvs, so its really a game of getting traffic, its hard to monitize and has almost no worth on amazon etc.
Frankly I hate amazon websites, still do, so this website had practically no affiliate etc, it was purely for ads and some rare posts where a product made sense (which never happened since I dropped this website before I ever put a single good link on it).

We are basically going to see where this goes from here, its moving from 150 human written posts, to 150 human written +3900 AI articles.

Will this go horribly wrong? probably.
Blame google for indexing 3k plus posts on a small website so fast lol.
crawl stats.
1677023558586.png

Crawl budget where?

This has only updated the first thousand or so post but beleive me when I saw all 3k plus are indexed.

1677023668704.png

its been going on 2 weeks and the website hasnt lost a single rank
Everything seems to just be slowly starting the uptrend, the lowest difficulty keywords have been falling into top ranks, while the hardest keywords in the niche are ranked position 50 plus. (which is to be expected).

Since the website didnt die the moment I did this, might as well share right? I dont really feel like these things will keep sustainable, so I dont really lose if this gets crazy or dies, since im the first person in my niche to do this.
The idea of AI taking up all SEO scares me because I make most of my living from SEO right now, but I might as well adapt if this is going to be a thing.
1677024307073.png
Yellow mark in image is roughly where the AI posts started, I will no longer be adding more posts since the whole niche is covered, I will only have a random post once a week to keep the website having activity.

Frankly, 5 days after I started this someone made a website copying the idea, however they are using a brand new domain and havent covered the whole niche as far as I see, I find it interesting to look at their website as its this fresh domain trying to do the same thing as my website, but it barely has an about us page, or contact info etc, it clearly is fully AI, while my site started fully human so at first look, you cant really tell what is going on.

I am considering updating high-performing AI posts into fully human written as they come as someone I know had mild success doing that but I'm not sure if I care that much, I would like opinions on if I should do that or leave everything as pure out of the box AI.

I just want to say to anyone who somehow knows exactly what website is this, please just keep it quiet and watch it, I feel the need to say this since its very obvious when you see a website ranking for every term in an entire niche all of a sudden.
To my competitors if this actually works out...... sorry bro, im sure an algo will wipe me out.

Hopefully someone here finds this mildly interesting as much as I found it interesting. Lets see if thousands of AI articles litterally kill this legitimate website.
 
Ranking will be temporary without proper backlinks and on page optimization. Usually lasts a couple weeks at best before you start to see most fall off of serp. Some will be revived on its own but you need some sort of backlinks for the most part. Doubt you'll get a any penalty though for 4k posts.
 
Last edited:
The site itself has thousands of backlinks (given most are just general internet spam) from the years its been ranking, the niche is rather educational so posts attract links. I have no current plans to do link building to the AI pages but that is subject to change. If I get an algo penalty or not will be found out eventually. Thanks for your thoughts. Also, if anyone has any links to similar posts where thousands of AI articles are put on a site, id love to see them (especially if they have traffic screen shots).
Ranking will be temporary without proper backlinks and on page optimization. Usually lasts a couple weeks at best before you start to see most fall off of serp. Some will be revived on its own but you need some sort of backlinks for the most part. Doubt you'll get a any penalty though for 4k posts.

So, it seems im finally starting to get a visible increase beyond baseline. Or maybe its just a traffic spike that would go away in a day (I checked analytics and the extra traffic is from the AI articles so I doubt),
It is rather amusing to me how much rank 1 spots these articles already have gotten in a few days, though I've seen some rank one spots fall right back to page 2 so its nothing I'm going to consider if they arent staying in a place for at least 2 weeks.
1677091872686.png

Just putting analytics veiw as it better shows when my traffic is flat or rising, my website has ugly bumps because nobody cares about this niche on week ends.
1677092106787.png1677092156663.png
 
I'm curious to know if the rankings will stay.
I did a similar test with Reddit and it seems my rankings are gone... not sure TBH.
 
Can you provide details on these questions?

1) What is the traffic change after you put AI articles?
2) Typically, what is the average content length of these AI Posts?
3) Do they (articles) include images and outgoing links?
4) What is the domain authority and page authority of site?
 
could you make a post on how you did you keyword research? Or could you just tell us here? How many average words per article?
 
Last edited:
Lets see if thousands of AI articles litterally kill this legitimate website.
I don't see any reason for this to happen. It will grow in traffic for sure because your website already has authority.

I think it's a good idea to augment the high performing articles.

How did you group the keywords for each article? Was it manually?

I think this experiment will be a success.
 
Some hard work done, please update us on how it's will be going in the future with the site
 
I'm curious to know if the rankings will stay.
I did a similar test with Reddit and it seems my rankings are gone... not sure TBH.
Me aswell, or at least if they would even have natural progression. If I posted a normal human article, it would take weeks to rank then weeks to drop if not years, if these have extreme volatility in serps it would lead me to believe google really is treating them different, however it could be other factors (like me not adding images, or youtube content because no one has shown me anything able to do this at the scale I want yet - please let me know if anyone has anything that can have relevant videos and images auto added to thousands of posts).
Can you provide details on these questions?

1) What is the traffic change after you put AI articles?
2) Typically, what is the average content length of these AI Posts?
3) Do they (articles) include images and outgoing links?
4) What is the domain authority and page authority of site?
Ill try
1) I don't think this is worth putting into a number since significant time has. If I compare my last 2 weeks with the 2 weeks before (since less than 2 weeks have passed since this has started) then my traffic gain is 32.75% currently over the 2 weeks before. Keep in mind the site is coming out of its off-season which is why I don't want to compare it to December.
2) Post lengths all range between 600-1000 words, I was just going for scale, I have no idea how successful or painful this will end up so doubling the word count (doubling the cost of doing this) wasn't really something I planned to do.
3) no, except for one single subcategory of the website (only 200 of the ai posts) I have interlinked to themselves to see if content cluster is going to change anything for them vs the rest.
4) DA isn't very high, its only 34, however, that is completely natural, link-building was never really done on this site.
could you make a post on how you did you keyword research? Or could you just tell us here? How many average words per article?
This is very niche specific, I don't think this will work in many other niches that have too much niche only terms and are too big. For this niche, it was picked by a younger me starting websites because it was easy to search for. I will just give an example for another niche so you get the idea.
Lets say most the keywords in the niche start like this "How to build a "insert variable keyword here, house, dog house, bird bath, you name it"".
I had every single keyword that begins with "how to build a" scrapped (use whatever keyword tool you have, Ahrefs, sem rush, etc etc.
Then have them all sorted by traffic, even if something has 10 traffic in a whole month it was added since keyword tools are often wrong anyways.
Then use a tool that can cluster keywords BASED ON SERP results, not based on just having the same words. As google knows that "how to build a fence" and "how to build a wood fence" are practically the same thing and can rank for either term, but if you don't cluster these then you will end up making articles targeting the same terms even though they have different words.
like 800 words avg

I don't see any reason for this to happen. It will grow in traffic for sure because your website already has authority.

I think it's a good idea to augment the high performing articles.

How did you group the keywords for each article? Was it manually?

I think this experiment will be a success.
In SEO nothing is guaranteed, I've had sites that on paper were perfect flop before, be it algo, or whatever. So this site could have anything happen to it, doesn't necessarily mean the AI articles caused the issue but that will be for us to judge in the future.
Im considering to change high-performing articles into human-written well done ones but....... As this is a test where anything could happen to this site, and I don't current have a site of the exact niche I can recover the articles to if this one burns, I'm not sure I want to invest in that. At the very least I am going to try it with 10 or so if the website has a big traffic hit, we can see if the updated articles actually kept their ranks. If you can show me other similar experiments that have success lasting a while then I might consider being more proactive.
No one in their right mind is grouping keywords manually, I had to turn over 10000 keywords into clusters since hundreds are just different spellings or very similar to other keywords. Frankly, even though I'm in the niche, its like asking a gardener if they know every possible name for different plants. Im not one for endorsing any sort of tool but anything like lowfruits etc that clusters by serp similarity is decent enough IMO.
Some hard work done, please update us on how it's will be going in the future with the site
I will try not to forget lol, I more often lurk on this forum than actually post, so this is probably the first time I'm actively showing things from my own sites.
Use an interlinking plugin to really supercharge this.
Any recommendations? I'm not sure exactly how to super charge this, do you know something that could automatically interlink them all?
Do you rewrite your AI articles?
No, testing bare bones, raw AI, 100% detectable.
Every rewriter I've tried doesn't deliver acceptable results as far as reducing AI detectability (except maybe jasper), however I'm sure rewriting 3 million words isn't going to be cheap, and idk if anyone has a way for me to do that at scale. If a rewriter isn't going to make it undetectable to Originality then I don't even see the point in me trying. Fooling free tools doesn't make me hopeful to fool google if they ever decide to crack down on this.
Which, despite what most people say, I think they will. It is almost crazy to image you could literally remake the entire Investopedia website in 3 days with AI and have something like that rank, even if it would never beat the real website, even ranking 1% of the traffic sites like those get can make a single person millions. I cant see in my right mind how this would be sustainable, I just started this experiment and already someone else has started doing the same thing in the same exact niche and rank next to me on some keywords, if just 10 people do this, that's a whole serp of AI rubbish the posters have never even read.
Just being honest because I assure you no one who posts thousands of AI articles has read all their articles. Even of the thousands I've posted, only around 300 have I seen with my own eyes. Some, I only see them because they are getting impressions and traffic.
 
Last edited:
but if you don't cluster these then you will end up making articles targeting the same terms even though they have different words
this is one of my delimma in kw research......the sorting out of the same kw with the same meaning or intent. With hundreds of them its so hard to sort them all out so aometimes I end up making an article pretty much the same with another LOL
 
In SEO nothing is guaranteed
Ofc that is implied. But what I meant is that AI content is not a reason to have problems if your website already has some traffic, authority and is already accepted in an ad network. Nor posting 4000 posts should pose a problem, perhaps 100,000 would be a red flag.

Your example with investopedia is good because in that case google should penalize, since it's a competitive ymyl niche and generic AI content would be seen as thin.

But for a smaller low comp niche like yours it should work. Even if 10 do the same, most of them will use fresh domains/websites and they won't be able to outrank you so they'll fail and stop renewing the domain next year. If a website with more authority than yours does this, then yes, they'll outrank you, but that won't happen very often and you can still have top3 keywords and get more traffic and earnings than before the experiment.
 
Back
Top