Why AI content beats most content writers with Michał Suski
Where to listen
Meet Michał Suski. He's one of the co founders of Surfer SEO, and he's also the Head of Innovation. So Michał, I'm gonna pass it back to you and tell us what you do.
Thanks for having me. Really appreciate the opportunity to spread some information about Surfer SEO, about SEO in general.
What I do right now, I'm playing with data. I'm playing with some data scientists to figure out new stuff for surfer. That's, that's my role now and it's pretty amazing and exciting stuff.
I might have to ask you questions off the record on data because that's what we do as a company and. That's my role.
So I want to ask what is your story of how you got started into SEO?
This is by accident, completely by accident. I was working as a QA specialist and my project was terminated. And I didn't have a project in the software house. I was like, all right, let's, let's do something. And it turned out that a friend of mine is running an agency.
In the middle of nowhere in a village, 400 ish people. So he was like, like a magnet collecting all of the people who were around and had some, like, whatever expertise in digital anything really. So I was like, okay, let's give it a shot. I watched the Josh Pashinski 2016 prediction to something about SEO.
And I was like. It's cool. It's changing environment. Let's, let's give it a shot. And I joined as a PBN copywriter, you guess. So that's how it started for me. The SEO journey started from writing blogs about roofing.
Exciting stuff. Pretty common story. Yeah, but no, you're learning the fundamentals. I mean, PBNs are, I mean, they're everywhere and they're not really disappearing anytime soon, but that's, that's a pretty good foundational start.
That is, that is crazy moment because right after I joined we were like publishing a lot of content. I was writing like eight pieces a day just to, you know, fill it with content. And we were putting exact matching anchor text backlinks in there, of course. And... There was a, I don't know what was that specifically, but some sort of a change and our client's rankings dropped.
And then I received an order to remove these links and change the anchors. So they are different. It was like two months apart. So one, one day I produced ton of content and do this linking. And the other day I wake up, go to, go to the office and I'm asked to remove these links. So like. What the hell? And you know what?
Of course, the, the, the rankings recovered. So that was crazy time.
I'm sure all that experience kind of led you to Surfer. So I'll jump right into a little bit of a side topic, which is I looked at your LinkedIn profile and saw that you've got a unique skill set and experience that makes you perfect for creating this type of BI tool, that's Surfer.
Your skills are kind of more I would say like data and products. So how are, how would you say they're important to what you do today?
Oh, today. Yeah. It has like, we are having a massive shift towards basing our decisions on data, on feedback, on quantitative and qualitative feedback, tons of tons of interviews and stuff, especially today with the AI.
We collect like hundreds of feedbacks and CSATs and stuff, running experiments, figuring out at scale, what works better for the people, what is really wanted, and we know that 57 people wanted something. And it's more than 56 people wanting something else. So we built that feature instead of the other one.
At the very beginning, it was all about the gut feeling. It was all about our hands on experience in working on SEO with SEO clients. So that was like. Figuring out our own problems, scratching our own itch just to be more productive at work and more efficient so that the optimization actually moves the needle and stuff like that.
But right now with like 20,000 clients at Surfer, we are able to. Incorporate more and more data driven decisions to the, to the product creation than we used to.
So would you say like when you started Surfer you and your founders did what you have at the beginning change over time?
Like, I'm assuming as you got more feedback from users and you started to realize how powerful data was. Was what you started with very different than where you are today?
Actually, no because I managed to keep the environment around me that is still based on the gut feeling as we have multiple product departments right now.
And they are working on, on stuff that like the AI, the content that these are like the products that we already have, but, I try to still use the gut feeling, still use the feeling with many different things and tests, like hundreds hypothesis a month. Instead of doing like 100 interviews and then building something to satisfy them.
So I'm happy to be in a position where I don't really have to base every single thing. On data because customers want to tell you a breakthrough they want like they will tell you a faster horse instead of a car and you need part of the team building faster horses and also you need part of the team building a car so sparking new ideas that won't come from a regular client.
Yeah. So that's kind of, so for me, it didn't change, but for the company, it was a huge change and massive, and I'm super proud we found people who can do it because I can't and people can. And that's, that's really awesome.
I understand because we're going through the same thing. You ask your customers for one thing and they, they always will say, you know, make the, make the faster car.
But it's not always what they want. And sometimes data alone just won't help you get the insights. But then of course. Once you build those things, you're, you're going right back to data to verify or validate that your, your gut instinct was right.
There is something about customer interviews and it's a great tip.
I'm not sure from what book, but I will just tell it because it's super important for everyone who is doing customer research. When you ask someone like, how did you do how do you do keyword research? And then they start to tell you like, we go here and there, but that will be their ideal wished how they want to do it.
But if you ask them, how did you do keyword research for the last time for a specific project you were working on? And then only then they will tell you the truth. If it will be tied to the specific activity they performed four days ago, how was the last time you did it? It wasn't like the perfect SOP executed one by one, you know, so to, to reveal the real real use cases, you need to ask for the real questions like, How did you do it for the last time, specifically, not in general, how do you do keyword research?
It's just the wrong question.
I agree. I jumped on a podcast not long ago, or I was listening to one and I heard a story from a guy named Tyler Ryan, who he shared on this podcast that he thinks that only 1 percent of affiliate programs are leveraging data. Like basically even looking at data to go, how can I make improvements?
So, I believe this to be true, and would you say this is similar in the world of SEO?
I hope it's going in the right direction, right? But, and hopefully, Surfer is a part of that change. But I, I, I must agree. Like if we consider SEO as a broad general niche of people having websites that receive some traffic from Google and they care about the traffic and we call these people SEOs, then definitely 1 percent maybe even less is using data.
But if we consider like people from like the Affiliate Lab, Maddigities group, then, well, definitely it's more than 1%. But like looking at the majority of people interested in organic traffic in general. Then it will be much, much lower number. Definitely.
Jumping into features about Surfer SEO, what would you say are some of the more underrated or underutilized features that your app has?
Underutilized features? I would say it's is the AI generator content. Because it's like, 25 ish, 30 percent adoption among the users right now. I understand not everybody wants AI content on their websites, but from what I observe the content is often better than human written content that is not super expensive.
Like, if you, if you compare, like, 50 bucks article from like a mediocre writer. It will be probably worse and 100 percent less researched. Like imagine like we as humans, we cannot. Analyze that big amount of information as the A. I. Like, it's just impossible to read the whole book in five minutes and then use all facts and interesting talking points from that book to create a review of that specific book.
It is just impossible. So I would say it should be used by 50 plus percent of the people. Thank you. I'm, I'm doubted it should be, it should be this way. So even though it's a huge contributor to our growth recently, I would say it's still underutilized feature of Surfer.
I think what you said is true, because I mean, I've been in that situation where we've had writing teams for my affiliate sites and it's the same thing.
It's under research content. You're paying anywhere from 20 to 50 and you look at it and you're like, I think I should have written that article myself. And it's not easy just to say, well, I'm going to write all of my content. But I think there's something to be said there where most content produced is just probably not worth publishing.
Imagine a scaling up. Like you just discovered your new niche for your blog about cats. You were into food, but now you are into, into something else, like little houses, cat houses, or like, whatever, really. And you want to tackle it because you just discovered e commerce that will pay you 50 percent of the, like, 50 percent of their margin.
And you're like, let's do this, let's do this. What do you need? You need to establish a little bit. Of topical authority. How do you do it? You create a proper silo that contains 30 to 50 to whatever Your strategy is but in general it will be quite a mass Of content and what you have to do you have to call your writers and maybe within a month Or if you are super lucky and have a huge team within three weeks, you can receive the first drafts of your 50 articles, which is painful because you could be earning already the, the articles could be already settling down in the SERP, start getting traction.
You will have time to put internal links here and there and all kinds of stuff. While you are waiting for, for the, for the human writer's output, that again is not exactly how you wanted it. So it's not only slow, but quite like similar in value or sometimes even, even like below the expectations. So. It's just too much time wasted.
Okay, I'm going to ask a question that's probably a bit more self serving. So we do a lot of B2B affiliate content. I'd say similar to Matt Diggity where we do affiliate program reviews. And what we're trying to do to be different is capture a lot of data sources that we can't scrape from website. So it's like, does this affiliate program have postbacs, dynamic verbal tracking links?
How often is the data updated? Is it real time, every hour? And sometimes it's data, like you're not going to scrape it. And AI probably won't tell me the answer. So how would you say, where do you get that data from? Everything it's we, we have an app, so we, we basically scrape that data. So we're able to change, to notice the change in detection.
We also go to the affiliate program software and ask them across all of your a hundred customers. Is it the same? And they're like, usually it is, right? It's updated every 48 minutes. So we have that as a data point. And we do think at some point, someone's going to start copying our data, but we want to be first to market with this data.
And we're kind of wondering, like, how do we leverage this content where we want to do 10, 000 affiliate program reviews? And that's a, that's a daunting task, but we also want to use our data to our advantage, where I do think that AI will help. Could help to a certain level, but it'll stop short of those unique data points.
And I think that's what we want to do for our content.
Recently we tackled this problem. I mean, it's not a problem. It's an opportunity as we allowed people to provide their own information. Like, is it something unique? Give it to us. If you give it to us, the AI can categorize it as a fact and then build narration around that fact.
Which is, which is super cool when it comes to providing something outside of the SERP or even outside of the internet. Like your personal experience with whatever, or your knowledge from a book that wasn't scanned. And you know, you cannot find it on, on, on like anywhere. So this kind of, this kind of data points that are custom added to AI brief make a huge difference when you want to deliver.
Like 110 percent of what you can find in CERP. Because there will be always a part of data that you have to cover just to be compliant with what's required. But you do the extra mile with the custom prompt. You feed you feed your own unique research. And the AI can incorporate that as it was scraped anywhere else.
So that's, that's kind of, that's kind of how we, how we do it recently. That's super fresh stuff, actually. So maybe not everybody knows about it. Okay.
Like I said, self serving. I think you're giving me a few things to do for the next couple months, but hopefully a couple other people might find that interesting where.
How do you take other data sets that don't exist and what does it look like? And I think people should be focused on going, if you can pull data that other people aren't talking about, that should be your edge.
Yeah, like social media content, that would be awesome for the reviews, brand mentions. God damn, imagine having like the brand monitoring tool, like brand 24 or whatever.
And pull like positive and negative sentiment about Surfer SEO showing up on Twitter. And you can collect that, feed it to the AI and like replace the article with like refreshed section about what people say, right? Yeah. Like, like we just figured out another use case, but if we spend an hour together brainstorming, we could probably find like too many.
And we'll solve the world's problems while we're at it. Of course. Yeah. So you're listed as head of innovation at Surfer SEO. Can you share what that entails?
So I got a team of data scientists and we, right now we are trying to solve the clustering problem from a different perspective. Like we are clustering the whole internet at once.
That's, that's the, that's the thing I'm like interested in right now the most, but besides the, besides that particular project, I try to be like the SEO consultant for surfer just to Help navigate through these ideas coming from the customers. So like, are they really helping SEO? If we go this path, will it, will it hurt or help?
So that's, that's, that's my role in there to spark some new ideas, to help leverage other people's ideas because feature ideas come from developers often, this is, this is super big this is a goldmine actually, because developers know the code. I don't know how it's like how the data is processed and often it's super often, it's not a rare case that I don't know that something is possible yet for them is like I can do it in no time.
I just didn't know. It makes sense. And I'm like. Holy shit. I wish I was in your head earlier, like a week ago. So that's, that's kind of a snapshot of, of of these things. I'm, I'm, I'm on basically being everywhere where something is happening to check whether. You know, we can, we can do something about it, whether it makes sense or not.
And second part to that question as a follow up, do you think more companies and even affiliate marketers should think about having some type of research lab or innovation division in their company?
I think so. I mean SEOs are researchers SEOs are innovators in general. Of course, there is a huge group of people who just follow a blueprints.
That's fine because they work, but there are many and probably most of the SEO, serious SEOs test stuff. What is it if not a research, if not innovating stuff, like you are trying new tool, you are innovating your process you are trying to link from tier four. That's an innovation to your process and you see, okay, it worked.
Should I replace, replace all my H3s with H2s, whatever. Like these kind of A, B tests in SEO are like part of the, of the day to day job. If you are on a certain level like if you have people to do the repetitive work. All your time should be dedicated to figuring out how to improve the processes by doing something out of the box.
So yes, definitely SEO's affiliates should be in that mindset because we've got like many revelations, but they come up six months after they stop working, you know? And you want to be the one who knows earlier, so you can leverage that before it stops to work. So, yeah, that's that. Definitely do it, guys.
Awesome. Back to AI, given that you've got quite a few AI tools. Integrated into Surfer how are you, are you using and leveraging AI tools today to automate
Workflow? Hmm, like for SEO or Surfer SEO or in general for the people or inside the application?
I don't want to say all, but let's, let's just start with like things that you use Like, you know, chat GPT and AI tools for, for your own work.
So in terms of the chat GPT, it's more like a curiosity for me and kind of ad hoc stuff to generate something quickly for me specific, like title of my presentation or sending invite for better testing email stuff that I'm not really good at. And GPT will do. Way better than me as I am not a native.
So it turns out much better than, than when I was trying to, you know, brain it for an hour, no, no sense to, to, to like, to be this way. And when it comes to doing SEO I'm leveraging the AI content. We even have, a author on the blog named Surfer AI, like right off the bat, it's not me, it's not my friend, it's not anybody else.
It's Surfer AI and... It is the author of some of our articles that are actually ranking pretty well. So I'm, I'm, I'm happy to continue that. Plus there is a competition on SEO Signals Lab right now that is about rhinoplasty in North Korea. Some sort of make made up keyword that SEOs are trying to tackle on parasite websites.
And I just, I just created the little silo on medium about the surgery and we'll see how it goes after 24 hours. I was third, then I got the index and then I was four and again, the index. So we'll see how it goes. But in general this is, this is where I, I, I leverage AI. I just generate a ton of arguments.
And last question, what do you see of the future of affiliate marketing as it intersects with business intelligence?
With business intelligence, the affiliate marketers and business intelligence, I would say That would be yeah, there is there is something in like from, from top of my head, I would say that we have finishes and we have different providers of the products.
And let's be honest in here. If someone else gives you a better offer. Or more compelling or something, something that you are able to sell easier then you will swap the product. So BI can definitely help in figuring out how you monetize and not only the whole website. But even particular page, like this page is monetized by Amazon, but with monitoring of like many different providers, you know, that I don't know, some other e commerce will give you better.
And your decision is based on data on research that you are swapping the product from the Amazon or a primary choice. You switch from Amazon to the other provider just to optimize. The visit from organic, from social, from like whatever source. So yeah. Optimization of the visit on the affiliate marketers page can be done by business intelligence and data.
And that's just one use case, but we could probably do more.
That's interesting because I think it's something we're actually trying to tackle on our side to say, well, yeah, because I think about it this way. This is what we have. We were like the QuickBooks for affiliates where we pull the data in from all the affiliate programs.
So our job is scraping or API. And the things that we want to highlight make easier. And this is why I love your tool, because I love what you guys do. You gamify the whole experience of writing content. And I love how you just analyze data. So we're trying to do the same for affiliate marketers to say, what is your best converting brand?
Because I believe from just my experience as an affiliate manager and an affiliate marketer, that a lot of affiliates don't always. Like sometimes they'll send more traffic to who makes them the most money. I'm like, but if it took five times more traffic to make just a little bit more than the next site, you obviously should swap them.
And I think a lot of people don't really know what they're missing out on, or they don't pay attention to conversion trends. Your earnings per click in theory is a flat line, but let's pretend there's a, let's say we're even promoting SEO tools, you know, surfers over here and you've got a couple of competitors.
Now, if you opt to optimize your landing page and you create a new landing page. We predict, we assume that 95 percent of the time it should increase. If it decreases, you want to know about this so you can fix it. And as affiliate marketer, you should, you should have a dashboard that tells you it before you have to actually analyze the data yourself.
And that's actually what we're working on. And I think with what the way you explained or how you answered the question, I think there's a rabbit hole of. I actually took some notes and I think I think your feedback was actually super helpful to me.
That would be awesome to, to know, like, let's say you've got Ahrefs and SEMrush, right?
And let's say that they still have affiliate programs, so you can still make money and blah, blah, blah. And let's say Ahrefs gives you 20, SEMrush gives you 25. So the quick decision would be, let's go with SEMrush because it's 25. But if you look deeper in the information, you will see that two times more people will actually buy AHF.
So even though the commission will be lower, you will make it up with the volume. And if I had access to such an information before I waste 100, 000 visits on my website, that will translate into five figures, six figures. That would be great because I would have the six figures in my pocket instead of somewhere else.
So that, that's, that kind of insights that are deeper would be awesome. Not only like how much is it, right? That's too short.
And we're trying to blend data where from most affiliate marketers, we kind of assume that they usually get most of their traffic from SEO, but it could be social. It could be video, which is still a form of SEO.
It could be paid. And basically they just want to see the left to right. Yeah. It's. Here's the traffic and then here's the conversions. Here's the end result. And you kind of want to be aware of like what you said. If if one's going to convert at double the rate, it makes more sense. But what if you're ranking like super strong for a keyword?
You don't want to also make a wrong decision because sometimes SEMrush, actually, if you're ranking, Like let's say you're ranked number two for SEMrush review. It's you're going to get high conversions from that. So that would be looking at data in the wrong way. And it's all about going can we connect all these sources and have it?
So it's blended.
Yeah. Even segmented by the, by the channel acquisition channel, because from Tik Tok, they would prefer some fancy shoe brands, however, from the organic. They will prefer Nike and Adidas because they're, they are like older people and they, they are into these. So you want to balance it, see like where the traffic is coming from.
And if it's worth to pursue the TikTok trend, go for it because. Their money is there. And if it's all about organic, don't try to be like too pleasing for the TikTokers because you don't have them. So go for the, for the standard stuff, Nike Adidas, like kind of, you know, that's made up example, of course, but this is how I see it.
In somewhere in there, I'm looking at the blue sky and thinking maybe.
Okay, last question and a bonus question. Do you think there could be a surfer for video?
So you're helping me the next ...
Well, it's, I'm using some video tools that will take insights. So for example, even this podcast, I'll put it into Opus. pro and it actually gives me the text and I can shorten it. Now I'm not going to change the text, but let's say if I wanted to create a video and I'm competing against Matt Diggity.
It's can maybe surfer is going to give me a great content editor to say, well, this is what you should create for ranking for this topic. And I think maybe there could be the next thing where I think video is going to become more powerful as. I mean, you're, you're helping change the whole internet where more AI content comes into play.
I do think people are going to go, maybe I want to listen to you on video rather than your actual content.
I would say that you just pointed me into like completely new stuff, but imagine imagine this with a recent app recent, almost recent update of YouTube providing transcripts of everything.
It turns out that they could actually do a text based algorithm, just like for websites, because they have transcripts, which is written content. So, if we could have, let's say, scrape top 10 videos on a keyword X, analyze the transcripts, and then, based on them, tell you the... Entities to mention topics like in general give you a script of the of the video that will be competing with these other videos based on text analysis of what they said, that's definitely a possible thing.
And if the video content is actually chasing. Organic from the text and this kind of stuff we do right now, then that will be a good opportunity to get into it, to tell you what to talk about in the video to rank. That's, thank you. I will make a note with my pen on my dirty something here, right?
Yeah, but seriously, this is good.
I'll expect that next year. . Awesome. Michał thank you for doing this. How can people get a hold of you?
SEO Surfers Facebook group or email is [email protected] .
Awesome. I'm in that group and it's a good one, worth joining. Thank you so much for doing this.
Have a good one. Cheers.