WEBVTT
1
00:00:00.000 --> 00:00:49.000
Hello, I'm Jeremy Rivera, your unscripted SEO podcast host, and I'm here with the legendary Patrick Stocks, who I've known quite a while now. I think most people in SEO should be familiar with them. If not, lift up the rock, guys. It's, you know, kind of nice out in the sunshine. He's been with Ahrefs for a while, but now he's striking out on his own and I'll add links to his stuff at the end. If anybody's looking for a great consultant, you'd be fool to pass up the opportunity to pick up Patrick. While I assume people know you, let's go back in time in history and talk a little bit about your past. Tell us where you've been and where you've gathered your expertise in SEO.
2
00:00:49.000 --> 00:01:17.000
Yeah, most recently was about six and a half years at Ahrefs. I was at IBM for like four years. Before that I was at an agency for a little while. I'm back to doing solo consultant, but I did that for a few years. Before that I was in house mid size B2B company. And long, long, long time ago I was actually a developer.
3
00:01:17.000 --> 00:01:27.000
Well, I know that you have those developer chops because, you know, I think at Ahrefs your official position was in the product manager seat. Is that correct?
4
00:01:27.000 --> 00:01:28.000
Yeah.
5
00:01:27.000 --> 00:01:35.000
And you, you built, you built some stuff. Tell us some of the stuff you built, man.
6
00:01:35.000 --> 00:02:27.000
I was looking at this the other day of like the, the product line and like so much of the, the new stuff or like my tools, my ideas, my roadmaps. But even when I joined like we were in the process of an infrastructure changeover. So if you remember when the version two became version three, I had a lot of influence on, on that, getting all the history and that kind stuff accessible and yeah, lots of charts, lots of reports, you know, page inspect, site structure, obviously huge impact on site audit. I should do one of those. Do you know the guy, what was it? Atlassian. He did that like breakdown of all the stuff he worked on. Should make a video like that. Very interest, but it might come off as like too bragging or something. I don't know. Might go off poorly, but a lot of, lot of things there.
7
00:02:27.000 --> 00:02:35.000
Is there a tool that you dreamed up that you didn't get a chance to build?
8
00:02:35.000 --> 00:03:25.000
Yeah, many. There's. I mean they still have a giant backlog over there of things, you know, I would say even ones that just never kind of like worked out the way that they should have. Page inspect, I always wanted the visualization, so we did A lot of like rendering and stuff. And basically the legal team was like, you can't show that. Oh, like we're storing the content. That's not great. Yeah, like that was. Well, some of that kind of made its way into firehose if you've checked out like firehose.com so having like the web history, being able to monitor anything on any page, that kind of stuff was, you know, it was supposed to be kind of a combination of like visual ping and like archive.org basically.
9
00:03:25.000 --> 00:04:35.000
From an SEO tools perspective, being part of Ahrefs, you have really gained that 800 pound gorilla reputation in the niche in the industry. At Ahrefs. Do you think that that's helpful or hurtful to the SEO industry overall? Because you have, you know, $100 price point between you and Semrush and anybody looking to develop a indie SEO tool, you immediately get compared to. Well, I could just go to a Ahrefs and is this really unique? Am I willing to. Should I, you know, just do ahrefs or does what you do in the separate indie SEO SaaS actually something unique and different? So I guess introspective question of being One of the two 800 pound gorillas, is that really good for the industry or does it pose some challenges?
10
00:04:35.000 --> 00:05:07.000
Both, I don't know, like, it's good to have a tool you can trust and rely on and at the same time, I think there is plenty of space for, for more niche tools. There's been a lot that like just came out, kind of carved their own space, keyword insights and tons of content tools and that kind of stuff that, you know, they can focus and do things better than the big ones. They can move a little faster at times. At least I hope there's space because I plan on building some things.
11
00:05:07.000 --> 00:05:50.000
Speaking of that space, what's it like sharing some of that space with. With a tool that also is going to be helping a lot of people fill that space? We've got AI capabilities and part of the dream that they sell is, hey, you can vibe code your own shit. But then that means, you know, we have more SaaS, capable people who just have this weird idea and they can get it to market faster. But then you also have that side play of those people have the budget, they can vibe code their own stuff. What does that interplay look like?
12
00:05:50.000 --> 00:07:18.000
Yeah, I mean, you can, but will the data be as good? You know, there are other data sources that are very much not as good as Ahrefs, for instance. So like, quality is going to matter. You know, people that have built their own tools know too. Like, there can be upkeep and all kinds of, like, issues that you gotta resolve. And realistically, like, if you're one person and you're suddenly responsible for 10, 15 tools and that's only part of your job, are you going to like, build as good of a tool as someone who has entire, like, product teams? Good question. I don't know that that will ever really, like, work out that well. Yeah, some people will build some very niche things and they'll get a lot of use out of that. You know, even Ahrefs though, like, they built Agent A and like, that actually was really cool internally, I was the fan. Cause, like, it was so hard for the product to mix like data sets or, you know, in the dashboard is kind of like, things have to be there in two seconds or people are like, pissed. So the responses had to just be there instantly. Now with AI mixed data sets, you could wait a few minutes to build stuff or like build again, like new databases, this sort of stuff, and then present it faster for you and that kind of thing. And I think it's a really, it's a, a really cool time to, to be building stuff.
13
00:07:18.000 --> 00:07:46.000
Do you think that, you know, as a SAS potential SaaS developer, that the broader acceptance of the beta test culture really helps, you know, meter those expectations when the big boys break their stuff consistently and it kind of makes it possible for you to succeed as a success as a SaaS with a little less polish?
14
00:07:46.000 --> 00:09:30.000
I mean, even, even Ahrefs, like the motto was like, first do it, then do it right, then do it better. So Ahrefs like broke stuff, but usually like, even like version ones of stuff sometimes are almost embarrassing to put out. I don't mean to sound that way. I love the tool, love the product, love the company people, everything. But it's like, it's, it's version one. It is, it's very much an mvp and then we do a big announcement. People try it and they're like, oh, that sucks. If they come back like three, six months later, it's awesome. So, you know, there's, there's a lot of that that I think people don't, don't realize that goes on. But yeah, smaller companies, yeah, you, you're going to break stuff too. And hopefully, you know, your customers forgive you and stick with you. It's just the reality of things when you're trying to do a lot or you're dealing with, you know, giant data and Ahrefs like the Infrastructure alone was like absolutely insane. Like, I have no question, I could never copy something like that. Like, no, no way. Even close. But. And you know, that means I'll never like get their data from crawling and that kind of stuff. So when I go and do stuff, it's gonna have to be a smaller scale, potentially a little worse data, or I have to, you know, potentially just do integrations, allow uploads of stuff because people are probably already going to have, you know, an Ahrefs or Semrush, they're going to have a screaming frog, they're going to have other tools that like, they're not going to go away from using those. They might use some of the things I built for other things, but like they're going to have their core tools still.
15
00:09:30.000 --> 00:09:34.000
Tell me about Tech SEO Connect.
16
00:09:34.000 --> 00:10:51.000
Yeah, that is our tech SEO conference. We are now in the third year. It's going to be November 5th and 6th, I want to say, in Durham, North Carolina. And it's awesome. We keep changing the venue. We make it hard on ourselves every year we bring different stuff. And yeah, it's great content, great people, lots of time for networking. We try to have amazing food, drinks and just some fun stuff. We had a couple themes last year of de stressing and then everyone on stage is talking about AI and automating everything and we had some real content creators in the space. We had this guy Chris, he's the poetry fox and he's like the poet laureate in Durham. He dresses in a giant fox costume. You give him a word, he bangs out a custom poem on the typewriter. Love every second of it. Yeah. And we had stuff like puppies just to like help manage stress and stuff because everyone's, you know, with AI especially, it's causing a lot of like bird out. It's a feels like it should be taking stuff off your plate, but it's change and then more to learn and more to do and more ideas and people are stressed.
17
00:10:51.000 --> 00:11:21.000
One of my last guests that said, I think it was Paul Pape, he said LLM and creating content at scale is now table stakes for small businesses, for across the industry. What then? If that's, you know, the baseline entry point, what is the frontier
18
00:11:21.000 --> 00:13:34.000
to actually make that content better, you know? Yeah, anyone can put out any number of pages. I'm literally like rebuilding my personal website and there's a whole like tech SEO hub of I don't know what it's up to 160 some pages or something right now. And it's all AI generated now, will it stay that way? The answer is no. Like, I actually will go through, I'm hoping I'll get some friends, like go through, add comments, tell me where I got stuff wrong. I have a feature actually where I'm going to have like users be able to report something's like inaccurate. But I did some stuff like summaries and, you know, checklists and like stuff that's actually like use useful and I don't see enough of that. Yeah, you can generate a bunch of stuff now, but if you're just using it for content, like, just, just the content, that's it. Like you're not doing a summary, you're not, you know, providing, you know, resources, links. Because these things like go out, they do the research, they create briefs. Well, you know, then if you're not like sourcing where you got something, what's the point? So, yeah, anyone can do it, but it's a matter of can you do it? Well, can you actually go. And, you know, now we've got systems that you could actually just have messages on Slack or emails or text messages, or have AI call people who are experts on every one of these topics and just do it once, twice, three times a week, whatever your content needs are. Message 5 People in your company that are actual experts that know these things or in your industry and like, have them, you know, answer some questions, get some insights, get some stories, experience, real experience. And I don't see people really doing that. They're, they're expecting, you know, this content that's generated to just work. And the reality is, you know, you've probably seen the mount AI stuff, it goes up, it goes down. And if you're not doing something different, if you're not doing something better, if you're not adding value, like that's absolutely going to happen, especially with informational content.
19
00:13:34.000 --> 00:14:01.000
What are some of the methods that you are turning to in your consultation to generate, you know, that unique information gain? You know, what are some workflows or processes that you're recommending? You know, instead of having somebody sit down and just put in prompts, is there some alternate process that you're embracing or encouraging?
20
00:14:01.000 --> 00:14:46.000
Yeah, again, I talked about like reaching out to the experts already, but leveraging any unique data that you have. Generally, I wouldn't even recommend a programmatic project unless, like you've got something unique, different, you know, if you're, if you're a directory, well, you've got reviews or whatever, everyone's got reviews. But like, what if you did scoring what if you did like a market analysis, like, there's so much opportunity for every one of these types of projects to do something different. What's the sentiment? What is the brand search volume for each of these companies? I would kind of give you an idea of where they are in the market for the position. Why is no one doing that?
21
00:14:46.000 --> 00:15:02.000
What would you say is your hot take in the SEO industry right now? What's the most controversial thing that you'd put out there about SEO
22
00:15:02.000 --> 00:16:07.000
that I think SEO is part of geo. Like people are saying it's the same, but really, if you're going to influence these systems, it's more than SEO. SEOs can only do so much. If you've got a problem with your product, you know the easiest way to get the LLM or the AI search system to say something different, Go fix that. Because if people are complaining, planning on the Internet, that's what it's going to say. So it's business, it's branding. It's so much more like these. These things is like you actually have to, you know, be a better company, build a better product, improve your services because the things that people are constantly complaining about, SEO can't cover that up at a certain scale. You know, if you're a mom and pop a local business, maybe an SEO can work on that. But big companies, they don't have enough SEOs on their payroll to like, go and like, change the opinion of the Internet.
23
00:16:07.000 --> 00:17:48.000
That's a good point because I was following a thread this morning about LLM tools seeming to hold a grudge. Like, if you got a negative review and picked it up, it seemed to echo for a lot longer than, you know, any good one. But they dug into it a little bit more and realized that it's because that same problem occurred in again in another one of your reviews. And so, like, you know, the sentiment was there because the sentiment's there, you know, and it's kind of, you know, stepping away from that golden age of SEO 2010-2020 where Google was monopolistic, SEO was ascendant, and you didn't have to do anything other than open Ahrefs or Rush, crank out some articles, find some backlinks, fix some tech SEO problems, and you're square. You don't have to worry about rogue LLM models ingesting comments from Reddit and surfacing it. We are entering a much more competitive era. You know, you have genuine entries from Perplexity, Claude, Copilot, Gemini that their models operate differently functionally, you know, and there are things that you can do now specifically for Claude or specifically for GPT, that don't seem to move the needle on the other models and also have their own separate impact on SEO. So it's a bit of a wild, wild west and a bit of a scrum, but I'm actually kind of grateful. Are you?
24
00:17:48.000 --> 00:19:05.000
Yeah, it's, you know, they're going to go through the same learning curve like Google did. A lot of the stuff that, you know, works too well now, they're going to have to dial it back. I think we already saw that with Listicles and stuff, which I think about a year and a half ago, I was saying that's the first thing they're going to kill off. Like, no way. Anything that SEOs scale and start kind of spamming, they're going to dial it back. It might work now, it might hurt you in the future, it might hurt you for a long time if you really abuse these Systems. I think SEOs got lucky with Google. We got Penguin and stuff, and for a few years that was really painful. And then like, what was it, 2016ish? They're just like, oh, yeah, we forgive all your bad links. Like, no worries. They could have held that over us for another, like, 10, 15 years, easy. And these, these AI search systems, like, they may make that decision. So some of the stuff that I'm seeing companies do now, I'm like, that's obviously a bad decision. And the question is, how long will that hurt you? How long will they hold a grudge against you? So I think you gotta be a little careful on that kind of stuff too.
25
00:19:05.000 --> 00:19:40.000
I always find at conferences, the best SEO wisdom doesn't happen on the stage. No offense, but it's in the conversations at the bar, it's in the hallway afterwards where you find that weird anecdote from the guy sitting around with the snack, and you just start chatting about your stuff. In the hallways of your last tech SEO connect, what's something surprising? An anecdote or a story or a project that you were, like, genuinely struck by?
26
00:19:40.000 --> 00:19:51.000
You know, I don't want to share. I'm not, I'm not going to answer that. I don't want to share any detail. I got one advice and I'm like, no, because it's in my identified person or company, so I'm not going to answer.
27
00:19:51.000 --> 00:19:52.000
Fair enough.
28
00:19:52.000 --> 00:20:06.000
I will say there's tons of time for networking, though. There's tons of time. We usually, I think at one point we had tabs open at three different places. So There's a lot of bar talk. We have great food. It's great time come out.
29
00:20:06.000 --> 00:20:24.000
Fair enough. Keep your secrets and if you want them, go to the conference. Actually, that's good advice. In general, if you're trying to grow in the SEO space as a career, conferences are fantastic. But that does make me ask, kind of change topic a little bit.
30
00:20:24.000 --> 00:20:25.000
It.
31
00:20:24.000 --> 00:21:01.000
Let's talk about SEO as a career. You know, you've now you've been in house as a product manager, been at an agency, done freelance, worked in house on SEO. Kind of mirrors my own, you know, I worked with Raven as product manager of in house Freelance. Which one do you like the most and what would you recommend right now? If somebody's quote unquote, getting. Considering getting into SEO, which direction would you push them?
32
00:21:01.000 --> 00:22:51.000
Everything is different. I love my time at Ahrefs, honestly. Great company, great people, great product. Other thing, agency can lead to burnout. It's a lot of hustle and deadlines and a lot of stuff going on. In house is fun, but corporate and politics can be draining. Tons of meetings can be exhausting. Consulting has its own headaches too. Like, I like the freedom, but then there's, you know, contracts and business stuff and cash flow and all that. You know, if I, if I had to pick one, I really enjoyed ahrefs. But like, in house I think is usually better than agency. I do, I think I will enjoy the freedom on, on the freelance again. So it's really hard to just pick one and. But it's really hard to like, go freelance if you're not like, well known, if you don't have leads, that kind of thing. So you kind of have to build up. And that's gonna bring me to the other question of, you know, what, what should you do? And honestly, like, if you're trying to grow your career, just be active in the community, go to conferences, talk with people, go to local meetups if they exist in your area. They're kind of rare now, but just, yeah, go write. Like, I for years was just kind of quiet and doing my own thing. I learned a lot once I started participating in the community, started writing, started sharing, started speaking. I think that all made me a better SEO and all kind of led me into the position where I am. And it's a good position to be in
33
00:22:51.000 --> 00:23:16.000
anywhere online, you know, because I would say, you know, SEO had a super strong hub on Twitter and then there's a diaspora spreading out. So where are some of the different SEO communities where you've ended up and Found not just the spam, not just the self promo, but the actual like back and forth and useful conversations.
34
00:23:16.000 --> 00:24:03.000
Yeah, there are still like slack groups. Noah Lerner runs the SEO community. That's the big one right now. I still enjoy Reddit, you know, I mod tech SEO. So hopefully there's not too much like spam and nonsense there. But some of the others can be a little like messy but still entertaining. Yeah, it's everything split after Twitter went away. There's still, there's a good community I think on LinkedIn, of course. Still some on Twitter, some on I guess like Blue sky and threads. I don't know, I never actually joined either of those, but I know some people over there. Mastodon was a thing for a hot minute, but I feel like that's sort of dead. I could be wrong. Yeah, it's fractured right now.
35
00:24:03.000 --> 00:24:23.000
What do you think is the best practice for SEOs coming into the next year? Like what are some mandatory things that if you're in the SEO niche, you've gotta be doing this now?
36
00:24:23.000 --> 00:26:54.000
I would say experimenting with AI. Just try and automate any task that you're repeating. You know, I have been doing this kind of stuff for years because I have the development background. It was like 10 years ago probably. I wrote like a redirect matching script because I was tired of just like, hey, you know, this page should go over here, like hours and hours or days and days sometimes. And you know, I was, I was doing a lot of website migrations and I was like, come on, like there's gotta be a better way. And I looked in the machine learning stuff and it's like, oh, like I can actually like match content on this page with like content on this page. I'm like, oh, that's great, that's fantastic. So anything that you know is taking too much of your time that you're having to do repeatedly, go just try and you know, build a skill, try and improve it. Try and get it to where it's good enough that you're like, I can trust that. Or at least like 95% of the way there takes a lot of the burden off you and go build stuff. Everyone's got ideas for tools. I have a friend that I was just talking with him over a beer the other day and I was like, man, you gotta go like, try like build a site, you know, with, with Claude. Just, you know, use astro hosted on Cloudflare workers, etc. He's like, you know, my wife has had this idea and it was just, you know, A history of, like, the restaurant grades, basically because they're big foodies. So here's their sanitation scores over time. And he built it in like 25 minutes. And he was like, wow, like, this is a huge unlock. And it's just he had never tried to do it. He thought, like, you know, it's going to be so hard, it's going to be too difficult. He's not going to be able to do it. But really, it sounds like such bs, but, like, anyone can build anything at this point. Like, that is a real thing. You have an idea and you're like, oh, I don't know how to do this. I was doing some stuff on my own website and I was like, yeah, this is way beyond me for infrastructure and dev stuff and everything. And I'm like, hey, Claude, like, you just run me through this. And it's like, oh, yeah, go here, change this, run this. I'm like, okay, it works. Or if it doesn't work, it's like, okay, what went wrong? What do I need to actually do? And it'll run you through it. It's amazing.
37
00:26:54.000 --> 00:27:12.000
Claude, code, Claude, Desktop, ChatGPT, Grok. Different tools, different functions. Which do you like, which have you used and anything special that you've kind of unlocked?
38
00:27:12.000 --> 00:27:56.000
I've tried all of them. I'm, you know, I go between, I would say Claude and Jim and I right now. For a while though, it was all like chatgpt. Yeah, they. They kind of like surpass each other in certain things in the market. You know, I end up using Gemini and Nano Banana specifically for images because cloud doesn't really do image generation, but also just for a lot of, like, general informational knowledge stuff. But I mostly use. Well, if I'm coding websites, a clog code for sure. I use Cowork actually a lot now for, like, the file access and everything and just the regular chat. Also,
39
00:27:56.000 --> 00:28:16.000
I'm curious, what is the worst mistake that you've made using AI? What's something you did where you're like, oh, that was. That was bad. Or in retrospect, that could have been really bad, or you blew something up?
40
00:28:16.000 --> 00:28:45.000
Yeah, all of the above. I've broken stuff, like, so many times, like, hey, like, how do I fix this? Like, hey, can you, like, debug this? But one was just like, when I was first learning on, you know, the limits and all, I was using the API and I didn't know about, like, the caching or anything. And, yeah, I ended up spending, like, way more than I should have to do. Some things.
41
00:28:45.000 --> 00:29:00.000
What's, what's the state of play in backlinks these days? As important as ever? More important or just a sidecar to best practice?
42
00:29:00.000 --> 00:30:19.000
Yeah, well, I looked at this about a year ago with AHREFS data, maybe a little more than that. That could be like a year and a half. And, you know, the correlation with rankings went down overall compared to what we've seen traditionally. It was still high for certain things like local companies. And I forget what else at this point, but for like local SEO, they still matter. I think it was more than we had seen in the past. I would say, you know, for the AI search systems, they probably are not using this directly, but, you know, a lot of them are pulling rankings from like Google and Bing, who do still use them. So I'm gonna, I'll say, like, they're not using them, they're using them in a roundabout way, I would say, but I wouldn't be surprised. You know, the, the general trend I see in the industry is people are kind of going away from the, the links. And, and that means it's actually probably going to be more of a differentiator in the future as these systems start looking to say, like, what is a good result? Well, now they could look at the links and say now that they're more rare, the links actually matter more. Oh, go ahead.
43
00:30:19.000 --> 00:32:07.000
No, I absolutely agree. As it gets less of that shiny object syndrome as, you know, as the fewer outfits kind of make their. Can make their money, that ironically will flip it over to those that are still doing a good job on it. You know, not only are they going get the traditional good Google results, but they have a potential of any of these LLM models. Could. They could already be modeling, you know, some sort of PageRank algorithm. You know, that patent's been out for so long, and anybody that's worked on LLM systems, you know, knows that there's 25 different APIs available that give page rank type information. So. And with the amount of crawl that these LLM tools are dedicating to discover stuff, I mean, you guys at Ahrefs, you built your own index and built your own Google, so it's not out of the range of possibility for any of them to integrate PageRank explicitly into LLMs. And it seems crazy to say, like, Gemini is sitting right there within Google. It's not that crazy to think that some of the flags are scoring if they can within, you know, if they can integrate it into the ranking systems of Google results. It's not. It's that Much of a leap to say, hey, you know, is the Gemini team going have access to that same data that's being used?
44
00:32:07.000 --> 00:32:36.000
Yeah, yeah. And any of the like re ranker models or any of the like, you know, when you're choosing the citations, it's an easy data point to add. And even if it's not like a Google or Bing, they these other systems probably already have some access to the data. Again they may be building their own but I think we it's pretty common knowledge. A lot of them are pulling from common crawl and like they have their own version the metrics already there for the pages in the index.
45
00:32:36.000 --> 00:33:05.000
One of the biggest things I always learned listening to your stuff, reading your articles, listening to you speak, is your depth of understanding of the actual architecture of Google itself. Could you break out a little bit of what you find is useful to know about Google's infrastructure and system? Because that can help better inform your SEO decisions.
46
00:33:05.000 --> 00:35:33.000
So I actually have a whole deck, like a Hausearch works deck. I think it's on the Ahrefs slideshare that was built after an internal training. So when we were building the building. Yep.com, i was trying to explain this is what we know about Google systems, how they work and everything. This all public knowledge and all. There's a lot of really interesting stuff. The web rendering system alone is kind of a complicated thing. But what are data sources? You know, pick something small. What are data sources for these systems? And most people will tell you site maps, right? Or crawling. Well, there's more. RSS feeds are a thing, pinging the sites, the web hub subhub. What I think it's just web sub. Now there's a lot of different sources or ways of getting this information. And it's not that they're just going to like look at one or two things, you know, submissions in gsc, the indexing API. Well supposedly for a couple types of content at least. Any given system that you think about, there's a lot of nuance and things or when are they storing stuff, what are they storing at each level? All that can have some really interesting implications. The whole re ranking system, how all that works, works is kind of fascinating really. And a lot of that is just going to be like, okay, so you get retrieved but maybe you don't actually end up ranking that well. And we see the same thing in the AI search system. So like you get a bunch of pages might be like 90 pages retrieved and then like 13 of them get cited or 30 of them get cited? Well, why were some retrieved but not cited? And like that's a, that's a really good question to look into. And it's usually like you are relevant enough but like you didn't quite say the right thing or in the right way. It didn't measure quite as well against the result. And so someone else got cited for that. Even though like you're there, you're in the ballpark and it's kind of like the, the low hanging fruit of the AI era is like you're there but you're not quite in that result set.
47
00:35:33.000 --> 00:35:42.000
What do you think really is the role of the HCU within Google?
48
00:35:42.000 --> 00:37:37.000
Ah, I mean it's supposed to be helpful content. Useful content. We, they added the extra E and eat and everything. The what experience? Because a lot of people, SEOs in particular were making a bunch of content about things that, that they don't really know about. They're not really the experts on, you know, they're regurgitating what's out there, maybe making a slightly better page than what's out there because you know, you combine information from different pages. But if you're not the real expert, if you're not the one with those insights, if you're not the one doing tests, I love like just the nerdiest, you know, guys and gals in different niches that are just doing random stuff, random tests on stuff. You know, the project farms, the, the vacuum wars. There's a guy, I watch his stuff for car detailing and like literally all he does is he, it's, it's not even fully detailing, it's literally the pressure washers. And he's like, oh no, the flow on this is bad. Like replace this part with this and like do this. And he's just so hardcore about it. Like you could tell this guy really knows what he's talking about. And I think that's the kind of folks that they want to elevate. You know, HCU hit a lot of sites, especially affiliates. It didn't hit all of them. You know, when I was, I always appreciate, I'm, you know, I have a cat, my wife is super into cats.cats.com they send off the food for like testing, lab testing to make sure it's safety, nutrition profile, like what's in there that shouldn't be or does it have what it's supposed to have? Like that's the kind of level that you have to be like if you really want to succeed with this.
49
00:37:37.000 --> 00:38:22.000
So is that on integration of writing style, showing off your Expertise taking the time to like include anecdotes. Because if you need a bullet point, quite frankly you can just turn to, you know, Claude. Claude's great at bullet points, terrible at anecdotes and it certainly has no unique anecdotes. It can't, it can't have its own experiences. So does that come down to just writing style or is it also, you know, the other pain point that SEOs experience are experience now is brand and you have to establish this brand entity. So is it those two, the additional of those two signals I don't think
50
00:38:22.000 --> 00:39:46.000
writing would get you. But so far, you know part of the thing that they mentioned I think with that update was this images, images and videos that show like you've got hands on with this product that you're talking about that you've got actual knowledge of it. Not you pulled the stock photo offline and from the manufacturer and then you wrote a bunch of stuff like I. That's just, that's, that's the level that is going to get hit that is not going to cut or cut it. And you know, I think some sites that actually were doing a good job got cut off caught up in this. You know there was one that I think Sam oh had done a video on of like this running shoe guy and like the guy knew what he was talking about. Like super knowledgeable. I thought his content was great but his competitors were like cutting the shoe in half and stuff to like see the materials it was made out of and everything. I'm like how do you compete with something like that? Like that is just this guy was, you know, he obviously like went to the gym and everything and like knew, knew exactly what he was talking about. He had real hands on expertise and should not have been hit I don't think. But compared to his competitors he was still I think a level below where they were
51
00:39:46.000 --> 00:39:54.000
hard to hear that you know you want to rank better but quite frankly the other guys are better.
52
00:39:54.000 --> 00:40:46.000
Yeah. And again like he I think is one that unfairly got caught up. There were a bunch of those. You saw some very public ones. Hellofresh. I think I just saw something that they're like back up with a massive increase like yay. And yeah, Google I think went went too far on a lot of, a lot of the things and hopefully they like rein it back in even more because there are a lot of deserving sites. Again, even someone in my personal opinion, even someone that just aggregates the info better and makes a better website better layout like that is still A result I want. So I think, yeah, their systems went a little too far.
53
00:40:46.000 --> 00:40:52.000
Is there any question that I didn't ask that you hoped that I would have asked?
54
00:40:52.000 --> 00:40:57.000
Honestly, I didn't know what you were going to ask me. It's unscripted. I love it.
55
00:40:57.000 --> 00:41:20.000
What, what's a challenge? What's an itch you're trying to scratch in the middle of your own back? If you can outsource a question or a challenge that you're trying to solve, what's. What are you spending time trying to solve right now in SEO? And I'll ask the next guest and see if they want to take a poke at it.
56
00:41:20.000 --> 00:41:35.000
I mean, my problem isn't really an SEO right now. It's getting used to the business stuff and contracts and all that stuff again. And it's just time consuming. All that stuff is kind of annoying to set up.
57
00:41:35.000 --> 00:42:37.000
That's fair enough. So, I mean, there's. I've done a couple of really good interviews. I'll add them to the show notes, but maybe like a, maybe do a round table and gather together some different people to talk about just the business infrastructure. If you're becoming an SEO consultant, what are the best practices? You know, establishing, you know, should you aim for, you know, a monthly retainer? Should you ask for all of that up front? You know, do you peg your retainer to like an estimate of hours per month or is it outcome based? Is there particular software that you use to align on that? Is there terms or terminology to add or to avoid when you're, you know, getting that contract signed? What's the best process of walking your client through those expectations? And, you know, what does expectation setting look like in SEO? So I think I'll gather.
58
00:42:37.000 --> 00:42:51.000
What do you write at the contract? How do you handle the sales, the leads, get rid of tire kickers? There's. It's been a process.
59
00:42:51.000 --> 00:43:31.000
I feel you. I definitely do. As a long term freelance contractor on and off over the years, I've definitely made some huge mistakes. You know, the biggest one is, you know, letting clients pay later. You know, you gotta at least get half upfront, if not all of it upfront. That was a painful lesson, but there's some others. I think I'll take this as a hint and put together a panel where we just talk about the mechanics of SEO consulting and best practices and put it all into one place.
60
00:43:31.000 --> 00:43:36.000
Yeah, that's awesome. I'll watch that episode.
61
00:43:36.000 --> 00:43:50.000
If people are interested in accessing your expertise, hiring you on to build something for them, build their brand, build their company. How can they connect with you and get started
62
00:43:50.000 --> 00:44:01.000
right now? I would say message me on like LinkedIn. It's probably the easiest way. Otherwise Patrickstocks.com is a work in progress and you'll be able to contact me there person.
63
00:44:01.000 --> 00:44:03.000
Well, I appreciate you stopping by.
64
00:44:03.000 --> 00:44:07.000
Yeah, thanks for having me.
1
00:00:00.000 --> 00:00:49.000
Hello, I'm Jeremy Rivera, your unscripted SEO podcast host, and I'm here with the legendary Patrick Stocks, who I've known quite a while now. I think most people in SEO should be familiar with them. If not, lift up the rock, guys. It's, you know, kind of nice out in the sunshine. He's been with Ahrefs for a while, but now he's striking out on his own and I'll add links to his stuff at the end. If anybody's looking for a great consultant, you'd be fool to pass up the opportunity to pick up Patrick. While I assume people know you, let's go back in time in history and talk a little bit about your past. Tell us where you've been and where you've gathered your expertise in SEO.
2
00:00:49.000 --> 00:01:17.000
Yeah, most recently was about six and a half years at Ahrefs. I was at IBM for like four years. Before that I was at an agency for a little while. I'm back to doing solo consultant, but I did that for a few years. Before that I was in house mid size B2B company. And long, long, long time ago I was actually a developer.
3
00:01:17.000 --> 00:01:27.000
Well, I know that you have those developer chops because, you know, I think at Ahrefs your official position was in the product manager seat. Is that correct?
4
00:01:27.000 --> 00:01:28.000
Yeah.
5
00:01:27.000 --> 00:01:35.000
And you, you built, you built some stuff. Tell us some of the stuff you built, man.
6
00:01:35.000 --> 00:02:27.000
I was looking at this the other day of like the, the product line and like so much of the, the new stuff or like my tools, my ideas, my roadmaps. But even when I joined like we were in the process of an infrastructure changeover. So if you remember when the version two became version three, I had a lot of influence on, on that, getting all the history and that kind stuff accessible and yeah, lots of charts, lots of reports, you know, page inspect, site structure, obviously huge impact on site audit. I should do one of those. Do you know the guy, what was it? Atlassian. He did that like breakdown of all the stuff he worked on. Should make a video like that. Very interest, but it might come off as like too bragging or something. I don't know. Might go off poorly, but a lot of, lot of things there.
7
00:02:27.000 --> 00:02:35.000
Is there a tool that you dreamed up that you didn't get a chance to build?
8
00:02:35.000 --> 00:03:25.000
Yeah, many. There's. I mean they still have a giant backlog over there of things, you know, I would say even ones that just never kind of like worked out the way that they should have. Page inspect, I always wanted the visualization, so we did A lot of like rendering and stuff. And basically the legal team was like, you can't show that. Oh, like we're storing the content. That's not great. Yeah, like that was. Well, some of that kind of made its way into firehose if you've checked out like firehose.com so having like the web history, being able to monitor anything on any page, that kind of stuff was, you know, it was supposed to be kind of a combination of like visual ping and like archive.org basically.
9
00:03:25.000 --> 00:04:35.000
From an SEO tools perspective, being part of Ahrefs, you have really gained that 800 pound gorilla reputation in the niche in the industry. At Ahrefs. Do you think that that's helpful or hurtful to the SEO industry overall? Because you have, you know, $100 price point between you and Semrush and anybody looking to develop a indie SEO tool, you immediately get compared to. Well, I could just go to a Ahrefs and is this really unique? Am I willing to. Should I, you know, just do ahrefs or does what you do in the separate indie SEO SaaS actually something unique and different? So I guess introspective question of being One of the two 800 pound gorillas, is that really good for the industry or does it pose some challenges?
10
00:04:35.000 --> 00:05:07.000
Both, I don't know, like, it's good to have a tool you can trust and rely on and at the same time, I think there is plenty of space for, for more niche tools. There's been a lot that like just came out, kind of carved their own space, keyword insights and tons of content tools and that kind of stuff that, you know, they can focus and do things better than the big ones. They can move a little faster at times. At least I hope there's space because I plan on building some things.
11
00:05:07.000 --> 00:05:50.000
Speaking of that space, what's it like sharing some of that space with. With a tool that also is going to be helping a lot of people fill that space? We've got AI capabilities and part of the dream that they sell is, hey, you can vibe code your own shit. But then that means, you know, we have more SaaS, capable people who just have this weird idea and they can get it to market faster. But then you also have that side play of those people have the budget, they can vibe code their own stuff. What does that interplay look like?
12
00:05:50.000 --> 00:07:18.000
Yeah, I mean, you can, but will the data be as good? You know, there are other data sources that are very much not as good as Ahrefs, for instance. So like, quality is going to matter. You know, people that have built their own tools know too. Like, there can be upkeep and all kinds of, like, issues that you gotta resolve. And realistically, like, if you're one person and you're suddenly responsible for 10, 15 tools and that's only part of your job, are you going to like, build as good of a tool as someone who has entire, like, product teams? Good question. I don't know that that will ever really, like, work out that well. Yeah, some people will build some very niche things and they'll get a lot of use out of that. You know, even Ahrefs though, like, they built Agent A and like, that actually was really cool internally, I was the fan. Cause, like, it was so hard for the product to mix like data sets or, you know, in the dashboard is kind of like, things have to be there in two seconds or people are like, pissed. So the responses had to just be there instantly. Now with AI mixed data sets, you could wait a few minutes to build stuff or like build again, like new databases, this sort of stuff, and then present it faster for you and that kind of thing. And I think it's a really, it's a, a really cool time to, to be building stuff.
13
00:07:18.000 --> 00:07:46.000
Do you think that, you know, as a SAS potential SaaS developer, that the broader acceptance of the beta test culture really helps, you know, meter those expectations when the big boys break their stuff consistently and it kind of makes it possible for you to succeed as a success as a SaaS with a little less polish?
14
00:07:46.000 --> 00:09:30.000
I mean, even, even Ahrefs, like the motto was like, first do it, then do it right, then do it better. So Ahrefs like broke stuff, but usually like, even like version ones of stuff sometimes are almost embarrassing to put out. I don't mean to sound that way. I love the tool, love the product, love the company people, everything. But it's like, it's, it's version one. It is, it's very much an mvp and then we do a big announcement. People try it and they're like, oh, that sucks. If they come back like three, six months later, it's awesome. So, you know, there's, there's a lot of that that I think people don't, don't realize that goes on. But yeah, smaller companies, yeah, you, you're going to break stuff too. And hopefully, you know, your customers forgive you and stick with you. It's just the reality of things when you're trying to do a lot or you're dealing with, you know, giant data and Ahrefs like the Infrastructure alone was like absolutely insane. Like, I have no question, I could never copy something like that. Like, no, no way. Even close. But. And you know, that means I'll never like get their data from crawling and that kind of stuff. So when I go and do stuff, it's gonna have to be a smaller scale, potentially a little worse data, or I have to, you know, potentially just do integrations, allow uploads of stuff because people are probably already going to have, you know, an Ahrefs or Semrush, they're going to have a screaming frog, they're going to have other tools that like, they're not going to go away from using those. They might use some of the things I built for other things, but like they're going to have their core tools still.
15
00:09:30.000 --> 00:09:34.000
Tell me about Tech SEO Connect.
16
00:09:34.000 --> 00:10:51.000
Yeah, that is our tech SEO conference. We are now in the third year. It's going to be November 5th and 6th, I want to say, in Durham, North Carolina. And it's awesome. We keep changing the venue. We make it hard on ourselves every year we bring different stuff. And yeah, it's great content, great people, lots of time for networking. We try to have amazing food, drinks and just some fun stuff. We had a couple themes last year of de stressing and then everyone on stage is talking about AI and automating everything and we had some real content creators in the space. We had this guy Chris, he's the poetry fox and he's like the poet laureate in Durham. He dresses in a giant fox costume. You give him a word, he bangs out a custom poem on the typewriter. Love every second of it. Yeah. And we had stuff like puppies just to like help manage stress and stuff because everyone's, you know, with AI especially, it's causing a lot of like bird out. It's a feels like it should be taking stuff off your plate, but it's change and then more to learn and more to do and more ideas and people are stressed.
17
00:10:51.000 --> 00:11:21.000
One of my last guests that said, I think it was Paul Pape, he said LLM and creating content at scale is now table stakes for small businesses, for across the industry. What then? If that's, you know, the baseline entry point, what is the frontier
18
00:11:21.000 --> 00:13:34.000
to actually make that content better, you know? Yeah, anyone can put out any number of pages. I'm literally like rebuilding my personal website and there's a whole like tech SEO hub of I don't know what it's up to 160 some pages or something right now. And it's all AI generated now, will it stay that way? The answer is no. Like, I actually will go through, I'm hoping I'll get some friends, like go through, add comments, tell me where I got stuff wrong. I have a feature actually where I'm going to have like users be able to report something's like inaccurate. But I did some stuff like summaries and, you know, checklists and like stuff that's actually like use useful and I don't see enough of that. Yeah, you can generate a bunch of stuff now, but if you're just using it for content, like, just, just the content, that's it. Like you're not doing a summary, you're not, you know, providing, you know, resources, links. Because these things like go out, they do the research, they create briefs. Well, you know, then if you're not like sourcing where you got something, what's the point? So, yeah, anyone can do it, but it's a matter of can you do it? Well, can you actually go. And, you know, now we've got systems that you could actually just have messages on Slack or emails or text messages, or have AI call people who are experts on every one of these topics and just do it once, twice, three times a week, whatever your content needs are. Message 5 People in your company that are actual experts that know these things or in your industry and like, have them, you know, answer some questions, get some insights, get some stories, experience, real experience. And I don't see people really doing that. They're, they're expecting, you know, this content that's generated to just work. And the reality is, you know, you've probably seen the mount AI stuff, it goes up, it goes down. And if you're not doing something different, if you're not doing something better, if you're not adding value, like that's absolutely going to happen, especially with informational content.
19
00:13:34.000 --> 00:14:01.000
What are some of the methods that you are turning to in your consultation to generate, you know, that unique information gain? You know, what are some workflows or processes that you're recommending? You know, instead of having somebody sit down and just put in prompts, is there some alternate process that you're embracing or encouraging?
20
00:14:01.000 --> 00:14:46.000
Yeah, again, I talked about like reaching out to the experts already, but leveraging any unique data that you have. Generally, I wouldn't even recommend a programmatic project unless, like you've got something unique, different, you know, if you're, if you're a directory, well, you've got reviews or whatever, everyone's got reviews. But like, what if you did scoring what if you did like a market analysis, like, there's so much opportunity for every one of these types of projects to do something different. What's the sentiment? What is the brand search volume for each of these companies? I would kind of give you an idea of where they are in the market for the position. Why is no one doing that?
21
00:14:46.000 --> 00:15:02.000
What would you say is your hot take in the SEO industry right now? What's the most controversial thing that you'd put out there about SEO
22
00:15:02.000 --> 00:16:07.000
that I think SEO is part of geo. Like people are saying it's the same, but really, if you're going to influence these systems, it's more than SEO. SEOs can only do so much. If you've got a problem with your product, you know the easiest way to get the LLM or the AI search system to say something different, Go fix that. Because if people are complaining, planning on the Internet, that's what it's going to say. So it's business, it's branding. It's so much more like these. These things is like you actually have to, you know, be a better company, build a better product, improve your services because the things that people are constantly complaining about, SEO can't cover that up at a certain scale. You know, if you're a mom and pop a local business, maybe an SEO can work on that. But big companies, they don't have enough SEOs on their payroll to like, go and like, change the opinion of the Internet.
23
00:16:07.000 --> 00:17:48.000
That's a good point because I was following a thread this morning about LLM tools seeming to hold a grudge. Like, if you got a negative review and picked it up, it seemed to echo for a lot longer than, you know, any good one. But they dug into it a little bit more and realized that it's because that same problem occurred in again in another one of your reviews. And so, like, you know, the sentiment was there because the sentiment's there, you know, and it's kind of, you know, stepping away from that golden age of SEO 2010-2020 where Google was monopolistic, SEO was ascendant, and you didn't have to do anything other than open Ahrefs or Rush, crank out some articles, find some backlinks, fix some tech SEO problems, and you're square. You don't have to worry about rogue LLM models ingesting comments from Reddit and surfacing it. We are entering a much more competitive era. You know, you have genuine entries from Perplexity, Claude, Copilot, Gemini that their models operate differently functionally, you know, and there are things that you can do now specifically for Claude or specifically for GPT, that don't seem to move the needle on the other models and also have their own separate impact on SEO. So it's a bit of a wild, wild west and a bit of a scrum, but I'm actually kind of grateful. Are you?
24
00:17:48.000 --> 00:19:05.000
Yeah, it's, you know, they're going to go through the same learning curve like Google did. A lot of the stuff that, you know, works too well now, they're going to have to dial it back. I think we already saw that with Listicles and stuff, which I think about a year and a half ago, I was saying that's the first thing they're going to kill off. Like, no way. Anything that SEOs scale and start kind of spamming, they're going to dial it back. It might work now, it might hurt you in the future, it might hurt you for a long time if you really abuse these Systems. I think SEOs got lucky with Google. We got Penguin and stuff, and for a few years that was really painful. And then like, what was it, 2016ish? They're just like, oh, yeah, we forgive all your bad links. Like, no worries. They could have held that over us for another, like, 10, 15 years, easy. And these, these AI search systems, like, they may make that decision. So some of the stuff that I'm seeing companies do now, I'm like, that's obviously a bad decision. And the question is, how long will that hurt you? How long will they hold a grudge against you? So I think you gotta be a little careful on that kind of stuff too.
25
00:19:05.000 --> 00:19:40.000
I always find at conferences, the best SEO wisdom doesn't happen on the stage. No offense, but it's in the conversations at the bar, it's in the hallway afterwards where you find that weird anecdote from the guy sitting around with the snack, and you just start chatting about your stuff. In the hallways of your last tech SEO connect, what's something surprising? An anecdote or a story or a project that you were, like, genuinely struck by?
26
00:19:40.000 --> 00:19:51.000
You know, I don't want to share. I'm not, I'm not going to answer that. I don't want to share any detail. I got one advice and I'm like, no, because it's in my identified person or company, so I'm not going to answer.
27
00:19:51.000 --> 00:19:52.000
Fair enough.
28
00:19:52.000 --> 00:20:06.000
I will say there's tons of time for networking, though. There's tons of time. We usually, I think at one point we had tabs open at three different places. So There's a lot of bar talk. We have great food. It's great time come out.
29
00:20:06.000 --> 00:20:24.000
Fair enough. Keep your secrets and if you want them, go to the conference. Actually, that's good advice. In general, if you're trying to grow in the SEO space as a career, conferences are fantastic. But that does make me ask, kind of change topic a little bit.
30
00:20:24.000 --> 00:20:25.000
It.
31
00:20:24.000 --> 00:21:01.000
Let's talk about SEO as a career. You know, you've now you've been in house as a product manager, been at an agency, done freelance, worked in house on SEO. Kind of mirrors my own, you know, I worked with Raven as product manager of in house Freelance. Which one do you like the most and what would you recommend right now? If somebody's quote unquote, getting. Considering getting into SEO, which direction would you push them?
32
00:21:01.000 --> 00:22:51.000
Everything is different. I love my time at Ahrefs, honestly. Great company, great people, great product. Other thing, agency can lead to burnout. It's a lot of hustle and deadlines and a lot of stuff going on. In house is fun, but corporate and politics can be draining. Tons of meetings can be exhausting. Consulting has its own headaches too. Like, I like the freedom, but then there's, you know, contracts and business stuff and cash flow and all that. You know, if I, if I had to pick one, I really enjoyed ahrefs. But like, in house I think is usually better than agency. I do, I think I will enjoy the freedom on, on the freelance again. So it's really hard to just pick one and. But it's really hard to like, go freelance if you're not like, well known, if you don't have leads, that kind of thing. So you kind of have to build up. And that's gonna bring me to the other question of, you know, what, what should you do? And honestly, like, if you're trying to grow your career, just be active in the community, go to conferences, talk with people, go to local meetups if they exist in your area. They're kind of rare now, but just, yeah, go write. Like, I for years was just kind of quiet and doing my own thing. I learned a lot once I started participating in the community, started writing, started sharing, started speaking. I think that all made me a better SEO and all kind of led me into the position where I am. And it's a good position to be in
33
00:22:51.000 --> 00:23:16.000
anywhere online, you know, because I would say, you know, SEO had a super strong hub on Twitter and then there's a diaspora spreading out. So where are some of the different SEO communities where you've ended up and Found not just the spam, not just the self promo, but the actual like back and forth and useful conversations.
34
00:23:16.000 --> 00:24:03.000
Yeah, there are still like slack groups. Noah Lerner runs the SEO community. That's the big one right now. I still enjoy Reddit, you know, I mod tech SEO. So hopefully there's not too much like spam and nonsense there. But some of the others can be a little like messy but still entertaining. Yeah, it's everything split after Twitter went away. There's still, there's a good community I think on LinkedIn, of course. Still some on Twitter, some on I guess like Blue sky and threads. I don't know, I never actually joined either of those, but I know some people over there. Mastodon was a thing for a hot minute, but I feel like that's sort of dead. I could be wrong. Yeah, it's fractured right now.
35
00:24:03.000 --> 00:24:23.000
What do you think is the best practice for SEOs coming into the next year? Like what are some mandatory things that if you're in the SEO niche, you've gotta be doing this now?
36
00:24:23.000 --> 00:26:54.000
I would say experimenting with AI. Just try and automate any task that you're repeating. You know, I have been doing this kind of stuff for years because I have the development background. It was like 10 years ago probably. I wrote like a redirect matching script because I was tired of just like, hey, you know, this page should go over here, like hours and hours or days and days sometimes. And you know, I was, I was doing a lot of website migrations and I was like, come on, like there's gotta be a better way. And I looked in the machine learning stuff and it's like, oh, like I can actually like match content on this page with like content on this page. I'm like, oh, that's great, that's fantastic. So anything that you know is taking too much of your time that you're having to do repeatedly, go just try and you know, build a skill, try and improve it. Try and get it to where it's good enough that you're like, I can trust that. Or at least like 95% of the way there takes a lot of the burden off you and go build stuff. Everyone's got ideas for tools. I have a friend that I was just talking with him over a beer the other day and I was like, man, you gotta go like, try like build a site, you know, with, with Claude. Just, you know, use astro hosted on Cloudflare workers, etc. He's like, you know, my wife has had this idea and it was just, you know, A history of, like, the restaurant grades, basically because they're big foodies. So here's their sanitation scores over time. And he built it in like 25 minutes. And he was like, wow, like, this is a huge unlock. And it's just he had never tried to do it. He thought, like, you know, it's going to be so hard, it's going to be too difficult. He's not going to be able to do it. But really, it sounds like such bs, but, like, anyone can build anything at this point. Like, that is a real thing. You have an idea and you're like, oh, I don't know how to do this. I was doing some stuff on my own website and I was like, yeah, this is way beyond me for infrastructure and dev stuff and everything. And I'm like, hey, Claude, like, you just run me through this. And it's like, oh, yeah, go here, change this, run this. I'm like, okay, it works. Or if it doesn't work, it's like, okay, what went wrong? What do I need to actually do? And it'll run you through it. It's amazing.
37
00:26:54.000 --> 00:27:12.000
Claude, code, Claude, Desktop, ChatGPT, Grok. Different tools, different functions. Which do you like, which have you used and anything special that you've kind of unlocked?
38
00:27:12.000 --> 00:27:56.000
I've tried all of them. I'm, you know, I go between, I would say Claude and Jim and I right now. For a while though, it was all like chatgpt. Yeah, they. They kind of like surpass each other in certain things in the market. You know, I end up using Gemini and Nano Banana specifically for images because cloud doesn't really do image generation, but also just for a lot of, like, general informational knowledge stuff. But I mostly use. Well, if I'm coding websites, a clog code for sure. I use Cowork actually a lot now for, like, the file access and everything and just the regular chat. Also,
39
00:27:56.000 --> 00:28:16.000
I'm curious, what is the worst mistake that you've made using AI? What's something you did where you're like, oh, that was. That was bad. Or in retrospect, that could have been really bad, or you blew something up?
40
00:28:16.000 --> 00:28:45.000
Yeah, all of the above. I've broken stuff, like, so many times, like, hey, like, how do I fix this? Like, hey, can you, like, debug this? But one was just like, when I was first learning on, you know, the limits and all, I was using the API and I didn't know about, like, the caching or anything. And, yeah, I ended up spending, like, way more than I should have to do. Some things.
41
00:28:45.000 --> 00:29:00.000
What's, what's the state of play in backlinks these days? As important as ever? More important or just a sidecar to best practice?
42
00:29:00.000 --> 00:30:19.000
Yeah, well, I looked at this about a year ago with AHREFS data, maybe a little more than that. That could be like a year and a half. And, you know, the correlation with rankings went down overall compared to what we've seen traditionally. It was still high for certain things like local companies. And I forget what else at this point, but for like local SEO, they still matter. I think it was more than we had seen in the past. I would say, you know, for the AI search systems, they probably are not using this directly, but, you know, a lot of them are pulling rankings from like Google and Bing, who do still use them. So I'm gonna, I'll say, like, they're not using them, they're using them in a roundabout way, I would say, but I wouldn't be surprised. You know, the, the general trend I see in the industry is people are kind of going away from the, the links. And, and that means it's actually probably going to be more of a differentiator in the future as these systems start looking to say, like, what is a good result? Well, now they could look at the links and say now that they're more rare, the links actually matter more. Oh, go ahead.
43
00:30:19.000 --> 00:32:07.000
No, I absolutely agree. As it gets less of that shiny object syndrome as, you know, as the fewer outfits kind of make their. Can make their money, that ironically will flip it over to those that are still doing a good job on it. You know, not only are they going get the traditional good Google results, but they have a potential of any of these LLM models. Could. They could already be modeling, you know, some sort of PageRank algorithm. You know, that patent's been out for so long, and anybody that's worked on LLM systems, you know, knows that there's 25 different APIs available that give page rank type information. So. And with the amount of crawl that these LLM tools are dedicating to discover stuff, I mean, you guys at Ahrefs, you built your own index and built your own Google, so it's not out of the range of possibility for any of them to integrate PageRank explicitly into LLMs. And it seems crazy to say, like, Gemini is sitting right there within Google. It's not that crazy to think that some of the flags are scoring if they can within, you know, if they can integrate it into the ranking systems of Google results. It's not. It's that Much of a leap to say, hey, you know, is the Gemini team going have access to that same data that's being used?
44
00:32:07.000 --> 00:32:36.000
Yeah, yeah. And any of the like re ranker models or any of the like, you know, when you're choosing the citations, it's an easy data point to add. And even if it's not like a Google or Bing, they these other systems probably already have some access to the data. Again they may be building their own but I think we it's pretty common knowledge. A lot of them are pulling from common crawl and like they have their own version the metrics already there for the pages in the index.
45
00:32:36.000 --> 00:33:05.000
One of the biggest things I always learned listening to your stuff, reading your articles, listening to you speak, is your depth of understanding of the actual architecture of Google itself. Could you break out a little bit of what you find is useful to know about Google's infrastructure and system? Because that can help better inform your SEO decisions.
46
00:33:05.000 --> 00:35:33.000
So I actually have a whole deck, like a Hausearch works deck. I think it's on the Ahrefs slideshare that was built after an internal training. So when we were building the building. Yep.com, i was trying to explain this is what we know about Google systems, how they work and everything. This all public knowledge and all. There's a lot of really interesting stuff. The web rendering system alone is kind of a complicated thing. But what are data sources? You know, pick something small. What are data sources for these systems? And most people will tell you site maps, right? Or crawling. Well, there's more. RSS feeds are a thing, pinging the sites, the web hub subhub. What I think it's just web sub. Now there's a lot of different sources or ways of getting this information. And it's not that they're just going to like look at one or two things, you know, submissions in gsc, the indexing API. Well supposedly for a couple types of content at least. Any given system that you think about, there's a lot of nuance and things or when are they storing stuff, what are they storing at each level? All that can have some really interesting implications. The whole re ranking system, how all that works, works is kind of fascinating really. And a lot of that is just going to be like, okay, so you get retrieved but maybe you don't actually end up ranking that well. And we see the same thing in the AI search system. So like you get a bunch of pages might be like 90 pages retrieved and then like 13 of them get cited or 30 of them get cited? Well, why were some retrieved but not cited? And like that's a, that's a really good question to look into. And it's usually like you are relevant enough but like you didn't quite say the right thing or in the right way. It didn't measure quite as well against the result. And so someone else got cited for that. Even though like you're there, you're in the ballpark and it's kind of like the, the low hanging fruit of the AI era is like you're there but you're not quite in that result set.
47
00:35:33.000 --> 00:35:42.000
What do you think really is the role of the HCU within Google?
48
00:35:42.000 --> 00:37:37.000
Ah, I mean it's supposed to be helpful content. Useful content. We, they added the extra E and eat and everything. The what experience? Because a lot of people, SEOs in particular were making a bunch of content about things that, that they don't really know about. They're not really the experts on, you know, they're regurgitating what's out there, maybe making a slightly better page than what's out there because you know, you combine information from different pages. But if you're not the real expert, if you're not the one with those insights, if you're not the one doing tests, I love like just the nerdiest, you know, guys and gals in different niches that are just doing random stuff, random tests on stuff. You know, the project farms, the, the vacuum wars. There's a guy, I watch his stuff for car detailing and like literally all he does is he, it's, it's not even fully detailing, it's literally the pressure washers. And he's like, oh no, the flow on this is bad. Like replace this part with this and like do this. And he's just so hardcore about it. Like you could tell this guy really knows what he's talking about. And I think that's the kind of folks that they want to elevate. You know, HCU hit a lot of sites, especially affiliates. It didn't hit all of them. You know, when I was, I always appreciate, I'm, you know, I have a cat, my wife is super into cats.cats.com they send off the food for like testing, lab testing to make sure it's safety, nutrition profile, like what's in there that shouldn't be or does it have what it's supposed to have? Like that's the kind of level that you have to be like if you really want to succeed with this.
49
00:37:37.000 --> 00:38:22.000
So is that on integration of writing style, showing off your Expertise taking the time to like include anecdotes. Because if you need a bullet point, quite frankly you can just turn to, you know, Claude. Claude's great at bullet points, terrible at anecdotes and it certainly has no unique anecdotes. It can't, it can't have its own experiences. So does that come down to just writing style or is it also, you know, the other pain point that SEOs experience are experience now is brand and you have to establish this brand entity. So is it those two, the additional of those two signals I don't think
50
00:38:22.000 --> 00:39:46.000
writing would get you. But so far, you know part of the thing that they mentioned I think with that update was this images, images and videos that show like you've got hands on with this product that you're talking about that you've got actual knowledge of it. Not you pulled the stock photo offline and from the manufacturer and then you wrote a bunch of stuff like I. That's just, that's, that's the level that is going to get hit that is not going to cut or cut it. And you know, I think some sites that actually were doing a good job got cut off caught up in this. You know there was one that I think Sam oh had done a video on of like this running shoe guy and like the guy knew what he was talking about. Like super knowledgeable. I thought his content was great but his competitors were like cutting the shoe in half and stuff to like see the materials it was made out of and everything. I'm like how do you compete with something like that? Like that is just this guy was, you know, he obviously like went to the gym and everything and like knew, knew exactly what he was talking about. He had real hands on expertise and should not have been hit I don't think. But compared to his competitors he was still I think a level below where they were
51
00:39:46.000 --> 00:39:54.000
hard to hear that you know you want to rank better but quite frankly the other guys are better.
52
00:39:54.000 --> 00:40:46.000
Yeah. And again like he I think is one that unfairly got caught up. There were a bunch of those. You saw some very public ones. Hellofresh. I think I just saw something that they're like back up with a massive increase like yay. And yeah, Google I think went went too far on a lot of, a lot of the things and hopefully they like rein it back in even more because there are a lot of deserving sites. Again, even someone in my personal opinion, even someone that just aggregates the info better and makes a better website better layout like that is still A result I want. So I think, yeah, their systems went a little too far.
53
00:40:46.000 --> 00:40:52.000
Is there any question that I didn't ask that you hoped that I would have asked?
54
00:40:52.000 --> 00:40:57.000
Honestly, I didn't know what you were going to ask me. It's unscripted. I love it.
55
00:40:57.000 --> 00:41:20.000
What, what's a challenge? What's an itch you're trying to scratch in the middle of your own back? If you can outsource a question or a challenge that you're trying to solve, what's. What are you spending time trying to solve right now in SEO? And I'll ask the next guest and see if they want to take a poke at it.
56
00:41:20.000 --> 00:41:35.000
I mean, my problem isn't really an SEO right now. It's getting used to the business stuff and contracts and all that stuff again. And it's just time consuming. All that stuff is kind of annoying to set up.
57
00:41:35.000 --> 00:42:37.000
That's fair enough. So, I mean, there's. I've done a couple of really good interviews. I'll add them to the show notes, but maybe like a, maybe do a round table and gather together some different people to talk about just the business infrastructure. If you're becoming an SEO consultant, what are the best practices? You know, establishing, you know, should you aim for, you know, a monthly retainer? Should you ask for all of that up front? You know, do you peg your retainer to like an estimate of hours per month or is it outcome based? Is there particular software that you use to align on that? Is there terms or terminology to add or to avoid when you're, you know, getting that contract signed? What's the best process of walking your client through those expectations? And, you know, what does expectation setting look like in SEO? So I think I'll gather.
58
00:42:37.000 --> 00:42:51.000
What do you write at the contract? How do you handle the sales, the leads, get rid of tire kickers? There's. It's been a process.
59
00:42:51.000 --> 00:43:31.000
I feel you. I definitely do. As a long term freelance contractor on and off over the years, I've definitely made some huge mistakes. You know, the biggest one is, you know, letting clients pay later. You know, you gotta at least get half upfront, if not all of it upfront. That was a painful lesson, but there's some others. I think I'll take this as a hint and put together a panel where we just talk about the mechanics of SEO consulting and best practices and put it all into one place.
60
00:43:31.000 --> 00:43:36.000
Yeah, that's awesome. I'll watch that episode.
61
00:43:36.000 --> 00:43:50.000
If people are interested in accessing your expertise, hiring you on to build something for them, build their brand, build their company. How can they connect with you and get started
62
00:43:50.000 --> 00:44:01.000
right now? I would say message me on like LinkedIn. It's probably the easiest way. Otherwise Patrickstocks.com is a work in progress and you'll be able to contact me there person.
63
00:44:01.000 --> 00:44:03.000
Well, I appreciate you stopping by.
64
00:44:03.000 --> 00:44:07.000
Yeah, thanks for having me.
