I Used ChatGPT to Rank #1 in Google (in One Hour)

231.15k views2737 WordsCopy TextShare
Ahrefs
I'm taking on a challenge to see if I can get to the top of Google's search results with AI-created ...
Video Transcript:
You might think that Google hates AI content because they have an entire section on it in their spam policies. But the keyword here that often gets missed is "spammy. " In fact, I think they love AI content when it's used in a very specific way.
I'm going to prove it by ranking #1 in Google, surpassing some of the biggest sites in the world using just AI content. In fact, I'm going to do it in less than an hour, starting. .
. now. Alright, so the way we're going to rank #1 is to go after a keyword where there's a Featured Snippet.
And the only real thing you need to qualify for a snippet is to rank in the top 10 for that keyword. So let's go and find those keywords. I logged into my free Ahrefs' Webmaster Tools account, went to the Organic Keywords Report, and set a couple of filters to find featured snippet opportunities.
And after a minute of work, I had over 7,000 keywords to choose from. Okay, so my game plan is to Google some of these keywords and choose one where I think the current snippet can be improved. I was going to make some minor edits to our post using content generated by ChatGPT.
I'm thinking like 15 minutes of work—max. And then from there, it's just a waiting game to see if Google thinks that our page deserves to rank #1. My hopes were high, and spirits were even higher.
But it was then that I remembered what happened the last time I tried this. . .
There's a hole in my plan. The last time I tried to make small edits and shortcut my way to #1, it didn't work. Basically, I had to rewrite an entire article in order to win that snippet, which, by the way, Google referred to as "epic tips.
" Anyway, I obviously don't have enough time to do a full rewrite, nor do I trust ChatGPT to do a full rewrite. More than anything else, I just really don't want to make the same mistake again. My stress levels were on the rise.
With time ticking away and no hope in sight, it felt like I was on the express train to Fail Town. But after a moment of success, two questions popped to mind that would set me on the right track. Okay… so what if last time actually wasn't a failure?
But rather, I didn't give Google enough time to give credit to those small changes I made? And what if. .
. what if rather than optimizing for one keyword, I did it for a whole bunch until my time ran out? I've got to try.
So back I went to the Organic Keywords Report to choose my keywords. Because there were 7,000 to choose from, I made a few rules for myself. Rule #1: Our blog post needs to be a good match for the query.
For example, this page ranking for "create an affiliate program" is a perfect match for the keyword. Our "Affiliate Marketing for Beginners" guide isn't. So, disqualified.
Rule #2: The page that owns the snippet has to have a similar or, ideally, higher level of authority than our website. That's because people love to comment with things like, "This is never going to work for 99% of websites. Ahrefs has one of the highest domain authorities.
" "It's worth mentioning that Ahrefs has high domain authority. " Yeah, we do. .
. but so do all of our competitors, including Wikipedia, which has a DR 96! And yeah.
. . I'm definitely going after Wikipedia.
We have a DR 91, in case you're wondering. Rule #3: Our pages can't have ranked recently in the featured snippet for an extended period of time. For PageRank, the last time we had and held the snippet was like 7 months ago.
This way, I'll know that a sudden #1 ranking for a sustained period of time would have been because of my work, not luck. So after 15 minutes, I had chosen 10 keywords to go after. It was time to call in my team of writers.
And by writers, I'm talking about ChatGPT. Alright, so the first keyword I have is "marketing funnel. " It looks like HotJar owns the snippet here, and it's clearly a definition or paragraph snippet.
So I'm actually going to copy this definition, go to ChatGPT, and say, "I want to win the featured snippet for the query 'marketing funnel. ' HotJar currently owns it with this definition. " And paste: "This is the definition from our page.
" I'll copy and paste that into our chat and say, "Now make ours better. " I couldn't believe the definition it gave me. .
. The definition is like a full blog post. No one wants to know this!
This is terrible. I thought this would be a home-run win. I was even planning to scale this to the 7,000-plus eligible keywords on our blog!
We'd get a vertical spike in traffic, I'd share it on our YouTube channel, and everyone would subscribe because they'd know that this is the place to learn digital marketing. But all my dreams shattered—from a single prompt. .
. and a single prompt is the very reason why this didn't work. Okay, so I remember reading in OpenAI's API documentation about system roles.
. . I can't find the exact snippet, but just look at their code samples.
They all define a system role like, "You are a helpful assistant," "a personal math tutor," or "an HR bot," and you have access to files to answer employee questions about company policies. Now, I'm not an AI expert or an engineer, but what I think a "system role" does is it adds context and expectations to the conversation. Like, what makes a featured snippet "good"?
ChatGPT and I have to align ourselves here so we're playing from the same playbook. Once we're on the same page, it's only then that I can ask for a snippet optimization and expect something maybe decent. So in an attempt to set context and expectations, I asked ChatGPT to tell me the different types of Google featured snippets.
The first two types were. . .
exactly what I wanted to optimize for. I decided to start with definition snippets, so I asked it: "What makes one definition snippet better than the other? " And ChatGPT outlined a checklist of criteria.
Now, because I agreed with most things said, I didn't ask for modifications or provide any further suggestions. I needed to start editing these posts because I was racing against time. So, I gave it my first target keyword, the content from the snippet, and their entire page's content for additional context.
I also told it my goals, gave it our full content, and asked why HotJars is better than ours based on the criteria we established before. And the justification was decent. Now it was the moment of truth.
"ChatGPT, can you rewrite my definition to be better than Hotjar's? Explain why it's better using our 7-point criteria as support. " And with only 35 minutes remaining, I saw something unfamiliar… hope.
Alright, so this still isn't perfect, but it's way better than what we had before, and it's actually working off a framework that I can reuse for all of the other keywords that I chose. So, I'm going to finetune this definition with ChatGPT, and then we're going to try and blitz through the rest of our keywords. We were off to the races.
I told ChatGPT to shorten our definition. I gave it some complimentary images that would be used in our post, and I asked it to use simpler language. And while it would have been way faster to just manually edit the definition myself, I was on a mission to prove that Google actually loves AI content when it's used to improve clarity and user experience.
Alright, so I've done the first featured snippet, and I feel like there's some chemistry now, and so, I'm going to go and try and finish off the rest of the 9 snippet optimizations. ChatGPT and I optimized snippets for the query "PageRank," "SEO tips," "SEO vs. SEM," and more.
We were like two peas in a pod. And with less than three minutes remaining, it was time to call it. Okay, so I did 9 out of 10 snippet optimizations with AI that I'm pretty happy with, so all we can do now is wait.
And we didn't have to wait long. With just one day passing by, it happened again… Yesss! !
I got my first snippet for "SEO vs. SEM"! !
! I was on an SEO high, and there was nothing that could bring me down. I went to bed thinking of what might happen over the weekend.
Will Google add me to their SEO starter guide page? Will they issue me a trillion dollars in stock? All I knew is that whatever was coming my way would be well deserved.
And before I knew it, it was Monday, December 11th. I logged into my Ahrefs account, went to my Rank Tracker Project, and it was gone… I lost the snippet. And the whole point of this wasn't to just win it, but it was to win and hold it for a sustained period of time.
So, we are back at square one with less than three minutes of our time. I need a break…. The next day came… then the next… and the next… but nothing.
I knew something was missing, but I couldn't quite put my finger on what. So, as a final attempt, I went back to Google and searched for "SEO vs. SEM" to see what could be missing.
Okay, so this is going to sound like a completely "conspiracy SEO"-type theory, but just stay with me here. So, I noticed that the SERP is still showing the old date. And that's because I didn't change the updated date in the post, considering it was a super minor edit.
So, I think I have like three minutes left from my original hour, so I'm going to just change the updated dates to the time the posts were changed. No bumping to the top of the blog or anything like that. Literally, just going to change the dates.
So, I did that with the 9 posts that I'd updated and submitted most of them to Google Search Console for indexing. My timebank had officially hit zero… And two days later, I'd learn that my conspiracy SEO theory may not have been as crazy as I had originally thought. Okay, so, it's December 15th and I think something's happening.
So, our share of voice had a noticeable bump for our featured snippet keywords – the most notable movement being for the keyword "PageRank. " And yeah, it's not a number one ranking, but I feel like something's brewing here. And as crazy as it might sound, I think it has to do with changing this tiny little part, which, by the way, has nothing to do with the quality of the changes.
Was I going crazy? It sure felt that way. So, I had to get to the bottom of this.
And I knew just the person to call to find out if I was crazy or if there was something happening under the surface. Okay, so, when I changed the definition for the PageRank article and Google recrawled the page, nothing really happened in terms of a boost in rankings. But when I changed the last updated date and nothing else, we jumped, like, 3 positions.
So, do you think that the date change had anything to do with this boost? Maybe. There's a lot of SEOs that believe if you update the date, it's going to help.
I think you probably triggered another system, maybe the Featured snippet system—you got a better answer and it's trying to rank you higher. Some of those processes are delayed a little bit, so I think that's a little more likely. So, you're saying there's still hope to win the snippet?
There's always hope. The fire inside of me was reignited. And I knew it was just a matter of time before we'd see success.
But with the holidays just around the corner, I decided to take some much-needed time off. I sang carols with the family, exchanged gifts, and rang in the New Year with a delicious burrito. And before I knew it, I was back in the office on January 3rd, checking our rankings… I did it!
! ! !
I can't believe I did it! Alright, so I got the snippet for "SEO vs. SEM," and I got the snippet for "keyword cannibalization"!
Also, we went from position 6 to position 3 for the keyword "PageRank. " So, we're holding it, and we're really not that far off from overtaking Wikipedia. But I quickly realized that the battle wasn't won because at least one of our keywords needed to hold the #1 ranking position for an extended period of time.
Days went by, and we lost the featured snippet for "keyword cannibalization. " But we were still going strong, holding the snippet for "SEO vs. SEM.
" And more importantly, our total share of voice was at a sustained new level. Something was happening. Then, on January 9th, the unthinkable happened.
Yooo, I just outranked Wikipedia! And by the end of the month, I hit peak adrenaline, winning 3 out of 9 featured snippets using just AI content. Now, as for the keyword "SEO vs.
SEM," we held the #1 position for well over a month, and the last time this happened was nearly a year ago. The same happened for "PageRank" with a tiny bit more volatility. By the time we were halfway through February, we were holding the snippet for 3 keywords: "marketing funnel," "SEO vs.
SEM," and "PageRank. " Okay, so we are at the end of February, and it's safe to say that this is a win, but I want to show you something even crazier than ranking #1 for a few keywords. So, I was tracking the 9 URLs I was working on in Ahrefs, and just look at the massive total traffic boost!
Even with just these tiny optimizations, we've got an estimated over 10,000 more monthly visits from search. This is amazing! !
! But then I realized that maybe I hadn't proven anything. I mean, did the ranking increase come from changing the dates?
Did we just luck out from some algorithm update? I knew that the only way to prove that Google loves AI content was to switch it back to the original human-written content. So, I went to archive.
org, pasted in one of our URLs, but they didn't have a snapshot of our page when I made the changes. So, I checked in Ahrefs Page Inspect. .
. Oh my goodness, there are so many dates to choose from. I chose the date right before I made the changes to this page.
I scrolled down and immediately saw the original version there. So, all I had to do was copy and paste this into WordPress and hit the Update button. Then I submitted it to Google Search Console.
Rinse and repeat for the other pages, and we’re back to the waiting game. But I didn't have to wait long… Alright, so it's only been like three hours since I submitted for indexing in GSC, and we already lost the snippet for "SEO vs. SEM.
" And let me remind you - we held that snippet for, like, two months straight… A few hours later, we'd lost the snippet for "marketing funnels" and "PageRank. " Now, while I may have proven that the AI content snippets beat our human-written snippets, it doesn't exactly mean that Google loves all content created by AI tools, clearly. Google loves what the user loves, and when a user searches for informational content, they want simple, clear, and concise information, which AI just so happens to be good at creating – probably better than humans.
So, rather than creating spammy AI content, risking getting deindexed and losing a large portion of your business in the name of scale… focus on clarity, and AI may just be a good assistant for that. If you want to see how to do that, go and watch our video where I ranked #1 for a super competitive query in under 24 hours.
Copyright © 2024. Made with ♥ in London by YTScribe.com