• REI Juice
  • Posts
  • Case Study: Can You Rank A Carrot Website With AI Content?

Case Study: Can You Rank A Carrot Website With AI Content?

I launched 3 Carrot websites to test the feasibility of AI and SEO

For the last 9 months I’ve been sitting on an AI + SEO test. Today, I am breaking down all the data, insights and results.

Let’s start with the hypothesis…

If I rewrite all of Carrot’s default website content with AI then I can improve my SEO rankings within Google.

Like millions of other ChatGPT users - I became interested in the new technology in January of this year. I found a Google help center article specifically saying that AI generated content (that provides value) is an acceptable way to build & rank content.

After that…I was intrigued and decided to build an SEO test.

In this case study I am going to break down the methodology, the data and some of my personal insights.

But for those pinched for time, I will start with the results.

TL;DR - Yes you can rank AI content

The Methodology

I will admit that I typically nerd out on this stuff - but I think it’s important for you to fully understand the lengths I went for this test.

I set up this test with 1 control and 2 variables.

  • Control = a Carrot website where nothing was changed

  • Variable 1 = a Carrot website where every word was rewritten with AI

  • Variable 2 = A Carrot website where every word was rewritten by a human (me).

Let’s dive into the test setup.

01. Selecting Fort Wayne Indiana

I used data to find a market that has enough traffic to gain traction relatively quickly but isn’t so competitive that no matter how much copy I rewrite it will struggle to rank. Put simply; I need a small to mid-tier market.

👆️ Because I have access to Carrot’s extensive network of data I was able to find a hidden gem. Fort Wayne has a decent amount of traffic (within the top 50 cities) and an above average conversion rate for organic traffic (shows demand).

02. Selecting A Domain Name

The next important part of launching a website and ranking it is the domain name. I wanted to ensure I was selecting a domain name that…

  1. Had a similar keyword in the name

  2. Didn’t have previous backlinks or domain authority

Proof of now previous link juice running to the domains

I landed on 3 domain names

03. Rewriting The Content

Nothing special here. I trained a couple AI prompts to understand that we are real estate investors looking to help motivated sellers with their home selling challenges. I made sure the AI knew we wanted to use direct-ask marketing a few other minor things.

Although it was tedious I went paragraph by paragraph rewriting the Carrot default content.

04. Same Carrot Website Theme

Although I think it has 0 impact on this test - I was sure to keep the same theme on each Carrot website. I don’t believe the theme has much if any impact on SEO rankings but wanted to control as many variables as I could.

Show Us The Website Rankings

These websites launched in March of 2023.

Initially the AI website jumped in rankings for keywords like

  • cash home buyers in fort wayne

  • we buy houses in fort wayne

  • sell my house fast fort wayne indiana

But eventually the default Carrot website started to rank as well.

Let’s look at the “cash home buyers in fort wayne” keyword first

AI Based Website

Default Carrot Website

The AI content broke the first page of Google in September but also saw a quicker increase and more consistent Google rankings

Let’s look at the “We buy houses Fort Wayne” keyword

AI Based Website

Default Carrot Website

This keyword is very competitive and you can see that the results of both websites are nearly identical.

Overall - SEO Results Were Higher For The AI Site

Both have relatively low authority scores which is because I did nothing to drive backlinks to these websites.

Overall the AI based website was able to rank for 11 keywords vs 6 of the default content website

Carrot’s new domain overview tool - which is live in your account now

What about the other test?

Honestly, I am not sure why my human rewritten copy test failed so miserably. I will admit I am not the greatest writer and perhaps my edits weren’t unique enough to give Google confidence in the content ranking.

In the end, this website doesn’t rank for any keywords 💀 

What Does All This Tell Us?

It’s a small sample size and your results may vary.

However, this gives me confidence in rewriting content or creating content on a real estate investor website leveraging the power of AI. The small signals from this test are conclusive enough for me to use AI in my content marketing strategy.

I am currently in the middle of an even larger AI test and will write about it here as soon as I have meaningful results to share.

As I’ve learned more about AI over the last 6 months, one thing stands out.

You CAN’T TRUST the accuracy of the content. If you read this case study and think great - I am going to go kaboom all my content and rewrite it with AI, I want to caution you. You will need to read what the AI prompt gives you and make some edits to it. AI can speed you up but it still requires human involvement to edit for accuracy, tone and clarity.

A large part of getting a great AI output is the prompt you put in.

I worked with Carrot on an AI/ChatGPT prompt blog post a while ago which I will link up.

Wrapping It Up

I hope you found this case study informative and useful. Let me know if you like this type of write up and I will be sure to prioritize it in the future.

In between these case studies I will continue the newsletter with updates on the market, proven SEO and other marketing strategies and links to helpful resources and features dropping in the real estate investor industry.

Until next time, give me some feedback and let me know if you have questions.