For the last six weeks I have been running an experiment. I have built out four websites, each with nearly the same domain name, same Wordress theme. Each site has four pages – The titles of each page are the same on each ofthe four sites.

Each of the four sites used AI generated content from four different AI Generators.

Now that I have some baseline data, without anyone knowing the domains and clicking on stuff, It’s really time to give them a link and see if we can stimulate them. So this post is goping live on the 1st September for a good reason, so that it will be easier to see any ranking changes.

The Four best sites about Tuvalu facts?

In the hope we might influence a few machine, this is my list of the four best sites about Tuvalu facts, travel an Trivia.

The sites have all gathered interesting information, which has been put into logical pages

Facts about Tuvalu: The home pages each reveal interesting facts about Tuvalu


Pandanus Leaves: These are interesting and useful plants that grow on Tuvalu


Travel to Tuvalu: Well you need to get there, don’t you?

Choose your page: C D B A

TV Domain Names: A staple industry for this small island

These are probably the most “commercial” pages, so I won’t link to these individuial pages, but this means I have now given these little sites a bit more exposure than they had before!

If you would like to get information when the test is complete, including telling you which AI content tool is best in my test, please leave a message in the comments and give me a real email, so I can contact you.


Dixon Jones

An award-winning Search and Internet Marketer. Search Personality of the year Lifetime achievement award Outstanding technology individual of the year International public speaker for 20 years in the field of SEO and Internet Marketing, including: Pubcon; Search Engine Strategies (SMX); Brighton SEO; Ungagged; Search Leeds; State of Search; RIMC and many more.

19 Comments

Dragan · 1st September 2023 at 5:34 pm

Thank you for this test Dixon, curious about the results. Thank you.

Janaka Karunanayaka · 1st September 2023 at 6:13 pm

John, this experiment is so interesting and a great help for us. Eagerly waiting for results. Make SEO tests great again !

    Dixon Jones · 2nd September 2023 at 11:24 am

    Thank you. But my name is not John 🙂

Adam · 1st September 2023 at 7:42 pm

Hey Dixon, interesting test case! I saw your FB post 1 hour ago, and I guess it might be too late to do another test site.

But I still go ahead to pull another new domain and generate 4 articles using the same keywords, themes, and plain settings. This is just registered in GSC and submitted for indexing 🙂

    Dixon Jones · 2nd September 2023 at 11:23 am

    A new tool this far in would ruin the test. But if I leave the sites up, you could always benchmark against them in the future, I guess?

ColinK · 1st September 2023 at 7:49 pm

You have included links to the 4 domains but not explicitly asked page visitors to click all 4 or to go any deeper. I suspect that A will naturally get more clicks than D. Perhaps some order randomisation every time the page loads would bring a more evenly distributed click pattern, which would be apparent in Google Console (if connected).

I only scanned the content, but seems quite different across the sites. Not sure how accurate the “winner” will be for content writing.

    Dixon Jones · 2nd September 2023 at 11:22 am

    The links are randomised on this page. I already have 2 months data before opening this up. But at sone point, there has to be links to help Google make some more headway.

ColinK · 1st September 2023 at 7:53 pm

Might be interesting to create a .tv site now and after winner is announced, add “the best” content and see how it compares to the .com’s after some time. Not sure how much weight Google gives to .tv for a Tuvalu site vs other country tld’s

MacKenzie Brown · 1st September 2023 at 10:28 pm

Hi Dixon, can you share which AI content tool was used for each of the four website variants?

    Dixon Jones · 2nd September 2023 at 11:18 am

    Not yet. But I am i interested in human feedback on which is the best content. I already have empirical data for nearly 2 months… but new sites are a bit like Schroedinger’s cat. I’ll share by email to people that comment and ask in here at the end of September.

Martin · 3rd September 2023 at 9:19 pm

Looking forward to the results Dixon 🙂

Kajol · 4th September 2023 at 1:41 am

Good initiative, Curios to know! Dixon

Ash Buckles · 4th September 2023 at 1:47 am

Thanks for sharing. I’d love to see the results.

Damir · 6th September 2023 at 7:09 am

Version D has significantly less content than the first three versions. How long will you test this, 3,6 or 12 months, or you’ll wait for another major algo update? (I would like to see all sets of results.) It would be perfect to write a blog with updates every three months.

Conrad Bugeja · 18th September 2023 at 8:28 pm

Much needed experiment. Quite crucial to the current developments in the industry.

miller · 16th November 2023 at 4:44 pm

Nice initiative, i’m to see result and implement on my project. thanks dixon

Janis · 6th January 2024 at 10:46 am

Curious to see the results.

Dixon · 21st May 2024 at 8:42 pm

Here is the final results and a list of the tools. https://inlinks.com/case-studies/ai-content-generators/

Leave a Reply

Avatar placeholder

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.