This website's SEO traffic went from 0 to 750K/mo SEO traffic using 100% AI-generated content

This content was created by Eric Siu. This content was uploaded by Eric Siu. This is a YouTube video (title: "This website's SEO traffic went from 0 to 750K/mo SEO traffic using 100% AI-generated content") published May 26, 2023. The excerpt is a single-speaker commentary/analysis about a viral SEO case study claiming a site reached ~750K monthly organic visitors using AI-generated content (notably Byword.ai). The speaker summarizes the case metrics (7,000 pages, rapid traffic and revenue growth, keyword rankings), explains the SEO methods (keyword research, topical pillar architecture, internal linking, crawl budget considerations, programmatic content production), and outlines a funnel approach (AI for top-of-funnel, humans for consideration/convert and fact-checking). They express skepticism about undisclosed conflicts of interest (suspecting the case author is a Byword co-founder), note indexing/backlink context and domain authority reliance, compare Byword to other tools (Jasper, ChatGPT, CopyAI), and discuss long-term white-hat vs short-term tactics and ROI. The speaker also references a follow-up thread, a Marketing School podcast segment with Neil, and promotes their agency SingleGrain and Byword.ai trial. Metadata: YouTube content, thumbnail provided, created/last edited 2023-05-26T19:00:01Z.
Chapters
This website went from 0 to 750,000 monthly visitors just from organic search using AI-generated content, and it only took a few months. First and foremost, my friend sent me a TikTok, and the TikTok actually went viral. It had a lot of likes on it, a lot of views on it, and then this Twitter post itself has almost 800,000 views on it, and it's not bad for a person that has only 3,000 followers or so. Okay, so this guy Jake Ward over here. So this is interesting. It's a great hook. So marketing 101, you have to start with the hook first. And you look at the stats here. So 7,000 total pages generated, hit $300K a month after 6 months, and it hit $750K after 12 months. And 4,000 keywords in positions 1 to 3, 13,000 keywords in positions 4 to 10. So positions 1 to 3 means rankings 1 through 3 on a search engine result page, also known as a SERP. And you can see this graph here. This guy shows that he isn't lying. So I think that's impressive. I think this is interesting. It keeps me as a reader hooked, and that's the job. It's to hook and keep them hooked, okay? To keep me interested. AIDA: attention, interest, desire, and action. So this website is in B2B SaaS, and someone actually reveals the— I think it's canopy.ai or canopy.app or something like that, but it's B2B SaaS, industry is finance, AI language model. So it Transitioned from GPT-3 to 3.5 to 4, and the tool I use is byword.ai. Survived all the recent Google updates. In fact, it thrived. Mainly, the key takeaway here is that he used this tool byword.ai, but I think we need to understand a little more on what he did. And there's actually a little frequently asked questions on this thread that he does a follow-up on. So understanding the target audience, a large part of a successful AI writer, what we research, keyword research tools. Okay. And that's nothing new. So if you want to grow from an SEO standpoint, you have to do your keyword research. You have to take a look at what keywords competitors are ranking for. And then you got to think about how you can create content that's better than it, also known as the skyscraper technique. And so he went for hundreds of potential keywords to target. Here are the 6 that he chose. And he focused around how to, Excel, define, calculate. My understanding is that he probably created a lot of different topical pillars. And so the idea here is that you start with one major pillar piece of content. Let's say it's how to get dressed for a wedding, right? And then your sub-contents might be, for men, for women, and then you might have different pieces on shoes to wear and things like that, and then they're all cross-linking to each other. So that's how you build topical authority when it comes to the game of SEO. And so byword.ai, he's claiming, can produce thousands of content pieces a day. And then you build that website architecture. So you have to create an area to house the content. And so what that means is the structure of your site matters because if you are not making it logical for a search engine spider to crawl where they see how things are organized, and how things are internally linked as well, also externally linked, then you're not going to be able to get the most juice out of your website. Because if your website is a complete mess when it comes to organization, you might be allocated a certain crawl budget because the search spiders are having difficulty getting through your site. So you want to make it really easy. You want to make it really logical. And even today's day and age, there's a lot more content that's being produced. And so you got to figure out how you can make your site stand out here. And so what's interesting is he claims that he did all this, but then this person tests the tool and he wasn't very impressed by the results. How long does cooked steak last in the fridge? So it should be cooked, grammatical issues. Cooked steak is a delicious and hearty meal. It's important to ensure that any leftovers are stored correctly to maintain quality. Explore the fact— okay, it doesn't even answer the question over here, okay? So now I have a little skepticism because if you can build to 750K in a span of 12 months or so, and that's monthly visits, I think that's basically 9 million visits a year. That's a lot of traffic. I've done SEO in the past. I've seen how these websites would work where I would make an affiliate site and I would drive the traffic to tens of thousands. And in one case, I got it up to 100,000 in a day. Very quickly, within a few weeks, Google noticed what I was doing and they torched the website, meaning that they burned the website down because I was experimenting a lot in my early days of doing SEO, all the tactics out there just to understand what was available. And trust me, it's much better to go for the long-term white hat strategies that will stand the test of time because ultimately it comes down to user experience and content experience at the end of the day. And so I think it's good that he achieved these results, but I'm a little skeptical at this point. So I think the next thing to take a look at here is that the follow-up thread over here. So 2 weeks ago I published my AI SEO study, got over 2 million views on Twitter and LinkedIn combined, hundreds of questions. Okay, here are the most common. So here's the analytics spam update, helpful content update. Okay, great. 90% sure that his results are legit. Byword didn't exist, so people are calling him out on it. Okay, so he's responding over here, but the way he's responding, it was a one-off consulting project for projects, but he's not saying full disclosure, this is my project. And so I was doing an episode on this on the Marketing School Podcast. Yesterday with Neil, and as we're doing, I'm like, you know, it kind of smells like he's one of the co-founders of the business. And so we continue doing the podcast, and we'll follow up on that. So will Google penalize the content? So Google is okay with AI-generated content, just not okay with thin content. Thin content meaning low-quality content. And so let's continue to go down here. Will Google penalize the content? What about editing and proofreading? Okay, so this is something I agree with. So we did virtually no editing at first, just published the content, then a periodic review once live, but only the pages that were ranking. Right now it's a gold rush to cover long-tail keywords with AI. Brands that sit and manually review each piece are going to lose out, the ones that are more aggressive. I kind of agree with that because even from a programmatic SEO standpoint, let's say we want to create a lot of agency pages on the Single Grain website. That's something we're looking at doing. If we're going to do that, it's hard for a human being to review 7,000 pieces of content, even 1,000 pieces of content. And so I think their approach to it is if we're gonna do 7,000, we're gonna put a human on the ones that are actually ranking. From a prioritization or workflow prioritization standpoint, that makes a lot of sense. ROI standpoint, there's a lot of traffic. It seems highly likely that it's ROI positive and we should keep doing it even if we can't Okay, the exact ROI, building presence in a broader target market. Okay, I kind of see where he's coming from. He wants to build the reach first before building the business, kind of how like a Logan Paul or MrBeast would build the audiences first and just continue to reinvest into building the audience. I can see where he's coming from there. And lower funnel content, is that possible? Okay, will AI replace content writers? To be clear, we still heavily invest in writers. So my take on this is that you can use AI writers, but it's still very helpful to have humans serve as the last mile. In terms of fact-checking a piece of content, upgrading a piece of content, adding links, adding citations, adding different examples, adding images, adding embeds such as rich media content such as video or audio. It makes a lot of sense to have a human being there. This part I actually found very helpful. So this funnel, which is what he uses, is an inverse pyramid here. You can call it a funnel as well. So it's awareness, consider, convert, similar to the AIDA model. So scale content production with AI for generic top funnel topics. That's the first step around awareness. And then the second one is on the consider piece. The majority of the time would use writers for personal experience, evoke emotion, and similar to below, probably only for the topics that are ranking or the articles or the blogs that are ranking. And then finally on the convert piece requires writers to explain nuances and ultimately convert readers. So the lower you go in the funnel, the more it makes sense to have the human beings at the bottom and then maybe some humans in the middle of it for the stuff that's actually ranking. And then you use AI content for top of the funnel, at least in today's standards. This could all change in 3 months or so given the speed of the AI news cycle. So we continue on. Okay, this is where I started to really question things. So Byword versus Jasper, CopyAI, ChatGPT. So if you want a writing assistant, a tool to help you write a few paragraphs or a rough draft, Jasper, ChatGPT, CopyAI, and the rest are perfect for that use case. So the way he's talking about it seems a little biased here. But if you want to generate full articles at a custom word count and a custom tone of voice with AI-generated images, with customized metafields, with a table of contents in multiple text formats, sync to your website CMS, and the ability to do this at scale, then use Byword. At this point, I'm like, he has to be a co-founder. And so Neil does the analysis here, basically searching, and we find out that he's one of the co-founders of Byword. So it's a little fishy that he didn't disclose that. He should have disclosed that. And maybe he's younger, I don't know. But I think it's still interesting that he published this case study. I think it's well done the way he marketed this. So let's continue on a little more. And so publishing velocity, majority of the 7K articles were published in January, February, 2023. Not going to lie, we were slightly nervous about indexing issues, but all the pages were indexed immediately. Unfortunately, we didn't have any issues. I don't have hard data, but I'll be slightly more concerned about fresh DA websites pumping AI content, especially if they're getting no traffic to begin with. In this case study, we leveraged the client's existing authority, natural backlinks from digital PR. And so let's take a look at this again. We didn't manually build any backlinks to website. As I said, we leveraged their existing domain authority. So domain authority scored on a scale of 1 to 100. The higher the better, and every ten points is basically a logarithmic jump, meaning that it's a significant jump each time. Like an eighty to a ninety is a huge jump. When you're above ninety or so, you're basically in the top echelon of websites on the internet. And so you can see the backlinks have gone up naturally over time as content was published. Then he puts you can try Byword for free. So he should have disclosed in even in his bio. He doesn't disclose that he's part of Byword AI. So I think this is interesting. I think it is worth testing out and go check it out. And by the way, hey, don't forget to subscribe. And go check out byword.ai. Let me know what you think you'd like me to cover next. And if you want help like this, we're covering a lot of stuff around AI marketing for your company. You can check out my ad agency, singlegrain.com, if you need help with things like marketing or even AI marketing. So that is it for today, and we'll catch you later.










