But creating 100 pages takes time and can be expensive.
As others have already mentioned, using AI (ChatGPT, GPT4, Midjourney, DALL·E 2 etc.) you can actually create a 100 page website far more easily now than before. However, if you just auto-generate the articles and post them all at once, it's more likely than not that Google will consider these articles as low value articles and not rank them well.
So, even though it became easier using AI, there will have to be some manual human input, scheduling over a longer period (e.g. no more than 3-4 articles a day), social media engagement etc. to make it valuable to your users.
If the content is interesting to real humans, it'll be interesting to Google as well.
Enter the minisite...
These sites are 5-10 pages long and center around a specific small niche.
Can these types of sites rank well for an uncompetitive keyword? The competitors for the keyword that I'm trying to rank for have DAs of 6-10. According the Ubersuggest, this is easy to beat.
Yes, mini sites can still rank. It's easier to rank well with bigger sites, but with good content, you can still rank with smaller sites. My personal sites and client sites do rank well for some long tail keywords. Regular social media activity will help as well.
Not anymore, you can generate a 1000 page site with ChatGPT in 15 minutes. There's even WordPress plugins you can use.
Yes, it's possible. However, as others have said before, these sort of auto generated content can be easily detected by Google. You'll need at least some human input to add value to those articles. Then you'll have to make social media engagements around those posts. Only then they'll rank, but rank they will if you can do it right.
One of the most prominent SEO ranking factor is Click Through Rate (CTR). Google often experiments with new articles on different sites and see how it does on CTR and bounce rate. If you add value to real users in your content, whether it's using AI or otherwise, you'll most likely get ranking on Google.
I am sure Google would like to / is working on detecting GPT-4 generated content. I think that is going to be a challenge, however.
There are already services available that can detect AI generated text and image content. I have tested some myself. They can detect even GPT-4 generated TEXT most of the time. You'll have to give some good input to ChatGPT and do some prompt engineering to avoid detection.
Example:
https://copyleaks.com/ai-content-detector
I'm sure Google internally uses better tools than these to detect AI generated content.
Google even heavily used ChatGPT API to train its own AI, as evident from OpenAI's CEO in this tweet:
So I'm certain that Google can detect ChatGPT (even GPT-4) output already, and they'll only get better at it.
Having said that, GPT by definition means Generative, and what it generates depends directly on user input. It can't generate by itself.
So, if someone can generate much better content than average people, and then put those content on small sites and if those sites become useful to at least some people - I don't see why Google will not rank them, even if they can detect them to be AI generated.
Why don't you start with 5-10 page mini-site and test your theory?
Yes absolutely, this is the best advice IMHO.
Don't depend on my opinion or anyone else's - do it yourself for a month or two.
Then you'll know for sure what works and what doesn't.
Happy domaining & content creation
Fayaz.
