- Niche Site Growth
- Posts
- AI content - yay or nay?
AI content - yay or nay?
Here's my take...
This might be a hot take right now considering Google’s recent crackdown on mass AI-generated content:
I think AI content is fine. As long as it’s helpful and accurate.
This isn’t just a blind opinion, by the way. It’s an opinion, yes, but I’ve looked through hundreds of sites, many of which are doing just fine despite heavy amounts of AI content. And I’ve seen plenty of sites with human-written content get absolutely destroyed.
How the content was first generated - whether from human hands or an LLM - is unimportant.
What’s important is the quality.
Things like:
Information gain
Positive user engagement signals
Writing for humans, not search engines
That last point is tricky. Actually, all of these points are tricky. Let me go through them one by one:
Information gain basically means “new information that can’t be found elsewhere on the internet”. It’s hard to achieve information gain with AI, because AI utilitizes already existing information when generating content.
But it’s not impossible. For example, you can combine two or more ideas or data sources to create totally new information.
Eric Lancheres gave a great example of this in his analysis of the March Core update:
With the advances in AI, creating fully automated sites with programmatic content has never been easier. The trick however, is to provide value (resources) while targeting keywords that no one else has targeted before.
Programmatic content usually combines different data sources to produce new content that is both useful and that has never been seen before.
For example, if you were to combine data from:
- Historical temperature records
- Current temperature records
- Umbrella sales
- Portable heaters sales
- Air conditioning units sales
You could create a programmatic website that publishes new keywords such as:
Weather-Driven Sales Trends [Date]
Weather-Driven Sales Trends April 2024
Weather-Driven Sales Trends May 2024
Weather-Driven Sales Trends June 2024
With new pages being published each month with pertinent data.
Weather Impact On Sales In [City / Region]
Weather Impact On Sales In New York
Weather Impact On Sales In Paris
Weather Impact On Sales In Montreal
This could be scaled by adding data for nearly every city / region that you have data for. You could easily create hundreds of thousands of pages with original, programmatic data that is useful to readers.
By creating new keywords at scale, your programmatic website could stand alone, capturing traffic from industry experts looking for unique combinations of data.
So yes, information gain is possible even with AI content.
Of course, it’s a lot easier to achieve true information gain with human-written content. Your unique knowledge and experiences are exclusively yours. When you share them online, that’s information gain.
In my research, I’ve found hundreds of sites in one niche that does particularly well through the March update. Can you guess the niche?
Recipes.
In particular, sites with real authors who share their own original recipes. These sites are chock full of tasty information gain!
Positive user engagement signals - things like time on page, scroll depth, clicks, etc. - are relied upon by Google to help determine intent match.
If the engagement signals are positive, search intent was likely matched and the user will likely remain a happy Google user.
The trick with user engagement signals is to satisfy intent, which you can do with either AI or human-written content. It’s more about how and whether you satisfy the user, not how the content was initially created.
Writing for humans, not search engines means publishing content your audience wants, not just targeting easy keywords willy-nilly.
This newsletter is actually a good example.
I write emails I think my audience will like. Sometimes people make a request and I’ll write an email about that subject.
I don’t care about writing for keywords, or only writing about topics that get a certain level of search volume. If I did, my emails would all be stuff like “how to start a blog”, “best SEO tools 2024”, or “plumbing SEO - how to get more plumbing leads from Google for FREE”.
…if you want me to write about those types of things, well then… I guess let me know? Lol.
Google can see what content is on your site.
If all your articles (or the vast majority) target keywords, it’s clear you’re just writing for search engine traffic.
That never used to be a problem, but it is now. Google wants to reward sites that build (or facilitate) a more personal connection with readers. That’s why Reddit, Quora, and forums are doing so well.
Try publishing more content that your audience would find super interesting, without worrying about keywords. You don’t have to ignore keywords and search volume altogether, though. A balanced approach is good.
Notice how this is about the subject of your articles, not the substance. You can publish endless amounts of AI or human content that’s not written for search engines. The content creation method doesn’t matter at all!
A word of caution
The draw of AI content is that it’s easy to instantly create gobs of it. Publishing hundreds or thousands of articles per day is far more attractive than spending all day crafting ONE high quality article.
If your goal is to rank in Google, I wouldn’t suggest publishing AI content without at least some level of human-editing.
Straight 1-click generated AI content can and does rank, yes, but most sites that did this were brutally hit by the March Core update. It’s a churn-and-burn strategy.
Put some thought into your content. Consider the markers of quality content I discussed above before hitting publish.
Thanks for reading,
Ian
What did you think of this email? |
Please give a rating to help my user engagement signals - Gmail uses them too! 😊
P.S. This newsletter is open for sponsors. If you have a relevant product or service you want to get in front of 6,600+ people (for a very reasonable price, I might add), you can see all the details, including pricing, here.