Pagination Best Practices For SEO & Google's Secret Weapon Against Spam | Mark Williams-Cook

Hello and welcome to this week's episode of Search with Kander. I am one of your hosts, Jack Chambers-Ward. I'm back in the Kander studio with Mark Williams-Cook.

Hello.

Hello. Welcome back.

Thank you.

It's nice to be, uh, cozy again. You know, we're getting used to this little setup in the corner here. We've even changed chairs. We were complaining about the chairs last time, so I've given you more back support, hopefully.

Yeah, I removed this.

Ah, okay.

I'm getting there. I feel a little bit too imperious.

I mean, you are the boss. Like, that's appropriate, right? It'd be weird if I was sat there and you were like— I don't know.

We'll flip it up next time. Or we'll both sit on the sofa.

Isn't that the whole reason we've moved out of the studio room is that we weren't sat next to each other? Anyway, we'll squish onto a 2-person sofa and do it that way. Anyway, we're not just here to complain about chairs like we did last time. We're actually going to be talking about pagination best practices for SEO and e-commerce and an interesting little Google patent all about catching out spammers, basically a little Google, Google sneaky little weapon to catch out spam tactics and things like that. But before we get to our conversation, you might have noticed we're both very blue today. We've got blue chairs, blue jeans, blue hoodies, because you might have noticed the logo. We've got Also Asked hoodies.

This merch came.

Actual merch for Also Asked. It's very, very cool and it's really nice embroidered. I don't know if you can see that on camera, but I'll put some photos and stuff in the show notes. They're really nice quality and I know a lot of people are going to be repping it at Brighton SEO.

So, so yeah, we, we took, we got an opportunity this year to do the Brighton SEO headline sponsorship.

Yeah.

Which is exciting. Thank you, Wix, for dropping out. Swept in there.

We're allowed in the building this time rather than spray painting outside, right?

Yeah, let's not talk about that. So Yeah, I wanted to do something because a load of the people that use Also Asked are going down, so I thought it'd be nice to get some hoodies printed. So I put out an offer like, if you're going to Brighton SEO, we'll get them done for you. I put an order in, and then when they arrived a couple of days ago, I was like, oh, they're here, I'll send them out. And then annoyingly, like, that post gets more reach than the original offer of free stuff. So then we've had like twice as many people ask for them and some people that aren't going. Um, so in the end what we've done, because we're raising money for Norfolk Clubhouse this year at Kander, we sorted it out so if anyone gives a donation to Norfolk Clubhouse of any amount you can afford, um, and you want a hoodie, then you can fill out the form and we'll send you one as well. So I'm going to be doing a second order.

Nice.

I was originally thinking at Brighton SEO to get people to wear them there because I thought it'd be really cool if we had like a bunch of people and also asking—

Like a flash mob almost.

Yeah, I was like, I was thinking about saying like if I see you in a Brighton— in an Also Asked hoodie at Brighton SEO, I'll buy you a drink.

How many hoodies have you ordered, mate? Okay, get dangerous.

But then I was like, if it takes 2 or 3 minutes at the bar to get— I'm just gonna spend the whole time like at the bar, even if it's just 30 people.

Did you really see the value from headline sponsor? Yeah, kind of. So I didn't do any networking, I just bought people beer.

Yeah, constantly. Um, so I just set up for the hoodies. in the end.

And yeah, probably a good idea.

Yeah.

Yeah. I'm excited to be a headline sponsor. That's really cool. I like the— it feels like, you know, we've done a little bit of sponsorship here and there and supported a few different conferences and stuff. Obviously featured at Also Asked at Search Norwich, the one that we run. And it's nice to feel like—

Women in Tech SEO, which is happening today.

We're recording at the time of recording. Yeah, absolutely. Yeah. It's cool to see us kind of like step up and be part of the the bigger picture, right?

Yes, it's one of those things I wanted to try because, um, while it was originally like an SEO tool, obviously it's good for anyone doing content.

Yeah.

But then when I take half a step outside of like the core SEO community, people are like, also what? Like, it's just never heard of it. And I— so it's an experiment to see how much impact that might have just having the name at, you know, Brighton's pretty much the biggest SEO conference going. So I'm interested to see how that goes. And it— again, I'm not expecting people to like sign up in droves. It's not about that.

Yeah, yeah.

Um, but yeah, it's interesting. It's cool to see. Um, and we'll see how it goes. It's cool being next to Ahrefs and stuff as well, like Some of the nice brands.

Definitely, definitely. So let's get into some SEO chat, shall we? And get into—

Yeah, I had something I wanted to railroad you first on.

Okay.

Which was super interesting. I saw you post about it today, actually, and I was going to post about it, but you beat me to it.

There we go.

Which was in our team, Tabby, who's our head of search.

Yeah, yeah.

Came up with a really nice point, which I hadn't actually really thought about.

Me too. Me too. That's why I wanted to highlight it. Yeah.

Yeah. Do you want to? Because I won't steal your thunder because you posted about it.

It's not my thunder, it's Tabby's thunder, to be clear.

I didn't want to steal Tabby's thunder. I'm leaving you to do that while she's at Women in Tech Festival.

Yeah, leave me to be the bad guy. Exactly. Yeah, yeah. The white guy steps in and just swoops up credibility.

So this is Tabby's thing.

For the record, Tabby's awesome.

And she's not here, so.

Yeah, yeah. I'm wanting to get Tabby on the show at some point. I'd love to get more of the Candid team involved. And Tabby is at the top of the list because she's actually come from in front of a camera. But anyway, Tabby mentioned in our chat recently, I think it was one of our clients had seen a bit of an influx in links coming in and they seemed, for want of a better phrase, natural. It wasn't like a competitor trying to spam us or anything like that, but there was a clear sudden shift. And Tabby pulled this brilliant, like, kind of connection and having a look at the actual content that is linking to us and the kind of things it's talking about. The theory is that a lot of people are using LLMs and stuff to write content. We know this, obviously. I know, shock horror, right? And the connection is that they are copying and pasting, including the sources and citations from LLMs into their content. So if you're featured as a cited source in an AI overview or Perplexity or ChatGPT search, whatever it is, you're going to get essentially accidentally included as a backlink from people's content. And that's definitely not something I'd thought of. I think it goes along, and I mentioned this in my LinkedIn post when I chatted to Tabby about it afterwards as well. There was kind of like, obviously the higher ranking sites and domains are in theory more trustworthy. So you're more likely to gain backlinks because you're more prominently featured, right? If you search for something and you're researching, oh, I should probably have a, you know, a source for this or another point of reference in this article I'm writing. You're going to search for it and find the top ranking things. Oh, that makes sense. I'll refer to this article. But doing that unknowingly, I think, is a really interesting angle where people will just copy and paste, not really paying attention to what they're doing from LLMs. And you get a bunch of accidental natural links. And yeah, it was not something I thought of because we always talk about the— it'll steal your traffic and there's no click-through rate despite Google saying more engaged than ever. Sure, Sundar. Sure. But there's that element of actually there might be a benefit, like a secondary tertiary benefit of more people unknowingly linking to your content if it does get kind of pulled up in those AI search things. What are your thoughts about it, Mark? Along similar kind of lines?

Yeah, I mean, we've had that effect always with when you rank well, you naturally pick up more links, which is kind of perverse really because you're already number one. So it's selfish.

You want to go and help the little guy out and just like, yeah, do some underdog linking.

Well, no, this, this is the conversation you have though, right? When you're like a newish site, you're basically, you're not discoverable.

Yeah.

So people aren't going to find you. And if journalists, for instance, don't find you, they're not going to link to you. So not only do you need to promote yourself so you're getting those links, you need to outstrip the people that are ranking. Um, and like you said, I don't recall seeing anyone directly kind of write about that before in terms of, oh, a benefit of being visible in LLMs is for when people are churning out content or even, um, using citations. Because I've seen some people using LLMs for that now because we know we can't trust the citations if someone's making like an argument or they're building a talk, they might say to especially a, um, a system like Perplexity, oh, find me some sites that do this. Like, so I use Perplexity the other day. So I've, I'm preparing a talk at the moment that's going to feature, um, some of the mistakes AI has been making. And I said to Perplexity, hey, find me some examples of funny things that AI has done wrong. And it went off and it searched a bunch of websites. And then I was checking those, but they're the links I'm gonna cite in my presentation, and they were found by an agent.

So if you're writing an article, that's exactly right.

It's an extra step. Yeah. So even— I don't think— the only thing I'd add to what you say is I don't even think it necessarily needs to be people copy and pasting, not checking. It's people being like, I'm writing an article about this, I was given these links, I've clicked them, they look legit, you know, they pass the sniff test of, yeah, that's basically saying that, and the site looks fine. I'll include it as a supporting— I'm not just making this stuff up. Um, yeah, and it'll be interesting to see how those 2 things interplay. Um, so I saw the study by Moz the other day saying that their DA metric was correlated— or it's a study actually done by someone else that Moz referenced— DA was correlated to how often brands and sites were mentioned in LLMs, which, which again seems kind of obvious because like you say, if they are prominent sites, they're likely to appear more often in, in the raw text training set. So I'll be interested to see how those 2 things interplay. But it was just a cool thing, I think, like you, where I was just like, oh yeah, yeah, yeah. And just made me feel good because we've got some smart people.

I know, right? I love that. Yeah, I always want to kind of— I know I've talked about this many times, but like shining a light on the awesome work that the Kanda team do that aren't, you know, you going to conferences or me being on the podcast, giving a highlight and a voice to the rest of the team. And like I said, I do plan to get some of the other members of the team involved in the podcast to talk about their specialist subjects and all that kind of stuff. So stay tuned for that, folks. Maybe you'll hear from Tabby, Brendan, Floor, and the team in the future.

So should we talk about what we were actually going to talk about?

Yeah, we're getting some Pagination. Yeah, pagination disagreements on LinkedIn. So this all spawned from an original post by Brodie Clark. Yes. Shout out to Brodie. He does amazing work with a lot of the SERP tracking features and all that kind of stuff. I live and die by his like record of what's changing in the SERPs and what new features Google's rolling out. Brodie does awesome work. I saw a talk of his a couple of weeks ago. It was really, really good as well. So shout out to Brodie. But he did start some, some interesting conversations, some heated debates in the comments. Um, we'll get to your comments as well, Mark, but it ended up kind of circling around the idea of pagination in e-commerce and whether the best practice of canonicalizing the paginated pages back to page 1 is the right thing to do in terms of what you see on the SERPs, in terms of user experience, is the traditional SEO best practice. And I've got the Google guideline quote here. Which is don't use the first page of a paginated sequence as the canonical page. Instead, give each page its own canonical URL. So that is the direct quote from the Google guidelines. Brodie was going against that and actually saying paginating back to page 1, sorry, canonicalizing back to page 1 can actually have its benefits in doing things like making your— the pages that are ranking for your site a bit tidier. So you're not getting a page 12 ranking for this particular thing. You just get a nice cleaner kind of URL. But I saw a lot of people disagreeing and I know Jan Karel jumped in and talked about splitting up different pages into— it's kind of like one big page but split up due to pagination. Of course, we had a lovely column come from John Mueller being classic.

Cryptic as usual.

Classic dry John of, I don't think a search engine needs to do anything special for pagination. Why shouldn't a URL be a URL? Classic John.

It's like a whole branch of SEO philosophy questions. Yeah, John, I get what he's doing in terms of the—

When is a URL a URL?

It's like a thought experiment rather than going into the technical minutiae of why things are how they are.

I remember last time we were at Brighton, you asked John for an SEO tip, right? And he just did some weird kind of cryptic, well, it depends what you mean by SEO tips or something.

Yeah, he gave me the, this is what I learned about B2B sales at Brighton SEO. Yeah, I think I caught him off guard there.

Yeah.

Sorry, John.

Um, but then, yeah, you, you jumped in the comments as well and kind of shared your opinions there.

I just thought it was an interesting thing to pick up on, right? Because it, I think it's a really good example of what I built my last talk around, which was that we have a whole bunch of SEO stuff we do, and there are some things that we know to be true and we can do things that affect those things. And then other things are like theory that we build up on. And sometimes I feel we get this, like, not colloquial, but like this village knowledge of how we understand SEO works.

Yeah.

And it's kind of agreed in the community that's how to do that. And it's built on a little bit of a false premise sometimes. And I really liked Brodie's kind of comments and post about it because he was actually really nuanced in his explanation. So, um, I think the key points of what Brodie was saying was, um, so acknowledge Google's best practice says don't do this. He acknowledged that there are large e-com sites that he has worked with that are doing this anyway.

It was like a lot of very high-ranking e-commerce sites just ignore this guideline and continue to rank really well. Yes.

Absolutely. He specifically mentioned things like product feeds and sitemaps being discovery mechanisms for Google, right? So, and this is, I think as well, where very specific use of language sometimes gets lost maybe in the general reading of what he's writing, because Brodie's been very specific there because he's thinking about paginated page sets in terms of, um, product page discovery. Yeah, not necessarily product page ranking, because discovery of a URL is very different to it ranking.

The discovery index ranking process that we talk about a lot, right? I know you talked about a lot of post-ranking stuff when we cover the exploit data a little while ago. Exactly.

But then he also specifically says, uh, when he's talking about pagination, he sees the benefit being not having indexed pages in Search Console that don't have the potential to rank.

Yep.

And this is maybe personally where I wouldn't say I disagree with him. I say I don't know if I agree with him because I feel through a lot of stuff that's happened with Google, we are now on this direction of Hmm, we should minimize indexed pages, crawl budget. Google's getting, uh, stricter with what they index.

We're gonna fight to be indexed, not fight to be ranked. It seems to be a common thing a lot of people are saying.

Yeah, so he hasn't— with the things he's mentioned, big e-com sites are ranking. I could say despite doing this.

Yeah, yeah.

Because big e-com sites are big e-com sites, and there's lots of reasons why big e-com sites will rank and can do what they like.

If you've ever seen a product description on Amazon, for example, it's an absolute shit show. Yes.

Um, so they can rank, you know, despite doing those things. Um, I also think that how an SEO feels about URLs being in or out of an index list that, you know, that at the end of the day it's just your feelings. It's like, yes, it feels tidier not having them in there. Does Google care? I don't know.

Yeah, yeah.

Um, so I was like, but it was a fair point. I was like, that's a really interesting point. And I picked up in this conversation because you say again, like, Jan kind of went in and was trying to get some specifics out of John Muir about like, what should we do? Like, you're not You know, it's not saying you're not giving us very clear advice here. Um, and then it's saying like a paginated page set is just one page spread over multiple pages. And then that's when I joined this conversation and say, well, Google are clear because they've got the guidelines.

Yeah.

That is their clear, this is what you should do. And this is the trap I think we've fallen into of, okay, if you think that a paginated page set is just one long page, It's literally not. It is literally a bunch of different URLs which are different pages. And I would say they're edge cases, but there may be cases where you search for— so let's give you— let's give an e-com example. We're selling, uh, we're selling boots and—

You didn't go for running shoes. I'm proud of you.

Okay, yeah, we're selling boots and we want, um, I don't know, black hiking boots. And on our site we only have categories for, say, hiking boots and, I don't know, building boots. That's not a thing, but you know what I mean.

Yeah, sure.

We don't have black hiking boots, but our pagination will say is alphabetically listed because it is. And on page 2, in the actual product description, we have black hiking boots. So we have maybe 10 products that are black hiking boots. Therefore, if someone did that search and Google decided your site should rank, your brand should rank, whatever, that second page of the pagination is actually—

More relevant, yeah.

The best page to land on. And for that to happen, it needs to just be a separate index page. Google doesn't need to do anything special with it. Um, and then we had this kind of discussion, this, this break-off, um, discussion around, well, those pages, the paginated pages, are all really similar. The only thing different to them is maybe the title or like some bit of this. Do I have to do different descriptive text?

It would often— you would often get like hiking boots page 1, hiking boots page 2. The only thing that changes in the title would be the number of the page, right?

Which, which is, which is fair enough. Um, I think, and I think that's helpful for the user. So when they see that, they understand it's a paginated page set. I don't think that's going to come as a shock. Someone's gonna be like, what does this mean?

Most people, even if they don't work in SEO, are aware of what Exactly.

So there's, I think there's a good user reason, reason to do that, that kind of tightening titling nomenclature. But again, what I think some people miss is the main content. So when we talk about quality guidelines, things like that, we speak at MC a lot, the main content is the product listings.

Yeah.

Like, I think sometimes again, we file away content in our head as text we have written people use this, but product listings are, are page content. So if you have a different set of products that are significantly different that could rank differently on page 2 to page 1, the main content's different. So it makes sense that both those pages should be indexed because they both have a chance to rank. There's obviously a bunch of other side stuff you can argue around, like, um, The reasons you don't block crawling, the reasons you don't noindexes. Google crawls all those pages and apart from discovering they do pass PageRank, they'll get those pages crawled more regularly. I'm always a fan of having things crawled more regularly than not.

Yeah.

It's not going to be any kind of downside to that. And I guess the final thing that came up there was, well, what about if we have like filters as well, like, you know, prices, things like that. And again, this is, I think, confused with pagination because they're separate things. Like ideally, if you have filters that produce URLs that you don't want indexed, they shouldn't be crawlable is like the best solution.

Yeah.

Like that's, I mean, in, in a perfectly designed site, you probably wouldn't need a robots.txt. Um, with e-commerce, it's there as kind of a bandaid of actually, can you not crawl these pages? You're going to get yourself in a bit of a, you know, a jumble if you do that. And I just thought it was a really interesting conversation that— and I really enjoyed the conversation as well because everyone was respectful and asking good questions and there was good back and forth. And John came in with his, why is a URL a URL? And then vanished. That was it. In classic John fashion. It ticked off, I've contributed to this thread, I'm out. But yeah, it was just an example, I think, of where maybe as SEOs, we've got an idea of this is how we think this thing should work. And then everyone seems to agree with me, and our logic is slightly faulty. I've asked Brody actually before this podcast because I didn't didn't have the chance to ask him before if in those cases where they have canonicalized back to the first page, if they are successfully getting those subsequent pages not indexed or their reporters not indexed.

Because we know canonicalizing doesn't guarantee anything.

Exactly.

Like I said, like you said, unless you're doing a noindex, which is, yeah, there's downsides to that.

Yeah, exactly. So I'm interested if that's, if that is uniformly the case. Um, because Jan did mention, oh, I had this issue with, you know, page 12 ranking, um, of this, you know, category page. And I have seen that happen before. Um, most of the time when that's happened, it's been because of internal links or external links. And so Google has decided this is the, the best page, which is annoying. And search engines aren't perfect. And sometimes, yeah, you do need to just kind of whack it with the wrench instead of using the wrench properly. Um, but I think they're the— they are the exception and making these kind of blanket rules because there are edge cases maybe isn't the best way. So again, I'm still— jury's out for me. And again, Brodie's specifically talking about e-com here.

Yeah.

So I think it's clearer cut for me when it's not e-com. Like, it's not e-com, I would strongly go with Google's advice. I think it's a little bit more edge casey with e-com where sometimes the main content of the product listings can be quite similar. And there, there might be confusion. But, um, I just thought it was a really interesting thing to discuss and, and go through because I was surprised by the amount of SEOs in there that were like, yeah, we do this.

Yeah.

And I would be surprised if any of them can definitively like prove that is the better way to do it because I'm all open for if you can do that, then yeah, let's do it that way. until Google fix it, because obviously if it's definitively the better way to do it, it means Google's probably not working as it should. Um, but it just seems risky to me too, because Google kind of know how they work. Well, yeah, not as well as they used to apparently. But, um, yeah.

Yeah, there's a, I think, a case I've seen where you see a paginated page ranking, and if the product— so sticking with the e-commerce example— the product range changes So I've had this where I've landed on take the page 12 as a random number. That's quite far into that product range. So chances are you've got dozens, if not maybe 100 or so products in that, assuming 10 products page, whatever it is. Then by the time you get to that and then they change, you mentioned the default alphabetical sorting or, oh, actually we should sort it by best-selling or featured or some custom sorting or whatever you want to do, or You discontinue or add a bunch of products that would then move the 10 that you would expect to see on page 12 to a different page. And like you said, oh, this is page 12 is actually where all the black hiking boots are, but you've discontinued 20 products. So now that's actually page 10. Yeah, because it's moved up.

Got it.

So you're kind of like, okay, you're stuck there and 12 is still ranking because you haven't come back around. You mentioned getting crawled more frequently and stuff like You haven't come back around and been crawled and re-ranked, for want of a better phrase. So you can get that kind of landing page where actually it was maybe the best option, but not anymore. If you've got a frequently changing dynamic product range, that might not be the best idea. It might be a good idea to have a more kind of stable URL set that is actually ranking more directly.

Yeah. So I think 2 things on that.

One, stop fiddling with your product ranges.

One, in all seriousness, like, one of the things Google has said they do in terms of crawl priority and frequency is they're looking at how often pages change. So my expectation would be if that's happening regularly, Google will probably crawl those pages more regularly to account for that.

Yeah, yeah.

It's interesting what you said about, you know, page 12, page 16, because that's the most common thing when we talk about what to do with pagination sets. People are like, well, we absolutely have to noindex or robots them past page 100. And it's like, why is there 100 pages of pagination?

Yeah. You don't split that into some subcategories? Like, that's— yeah.

That's a user experience issue. That's not an SEO thing.

Like, people are never going to find page 100. The whole joke of like, what are you going to do, bury a body on page 2 of Google? Like, the same is true for pagination. Nobody goes down to the bottom of the Google thing and clicks the second page there. And people are far less likely to click 99 pages through.

Oh, right.

Use an internal search, internal filter, something. I am not going to manually click through 99 pages.

I mean, the first thing I do on most sites that do that is products shown in this list is you put it on the maximum.

Yeah, right. Yeah.

Just so you're not doing that. Yeah. So, um, I would like 100 per page rather than 10. Yeah, I'd rather wait 3 seconds and just have it all on the page rather than having the cumulative interstitial loading time, which will probably be more, and it's more, you know, it's work, all that clicking. Um, so again, that, that I think that's like a— sometimes we get confused between this is a user experience issue, it's not actually an SEO issue. So we're trying to solve that issue with SEO tools isn't sometimes the best way to go about it. And I understand yet there are definitely going to be people working in enterprise here being like, yes, I know, I filed that ticket, had the 6 meetings about it. It was deprioritized and the person that was going to do it was laid off. And yeah, so we had to use robots.txt. Fine. But what I'm saying is they're Band-Aid solutions to, yeah, bigger, bigger problem.

Yeah, definitely. Well, let's get into something that was covered on the SEO Patent Podcast, the other podcast that we produce, quote unquote, not hosted by Mark and I, but in fact hosted by Delightful American robots. Yeah.

So if you don't know about that, um, it kind of started as a bit of a joke, but turned out to be quite helpful.

Uh, I believe you texted me while I was on holiday in Barcelona and said, I'm putting you out of a job. Can I have the Spotify login please? I was like, no, it was—

I don't think it quite went like that.

No, no, no. The other way around. Can I have this? What are the Spotify login details? And I'm like, I'm in a massive cathedral in Barcelona. I'm like, I don't know, Mark. Oh, let me just log in on my phone and see if I can make it work and give you admin permission. Why'd you need it? Just putting you out of a job, mate. Have a nice holiday. Like, oh, thanks, man.

So, um, yeah, I had for a while been using NotebookLM to generate discussions around patents that I uploaded from Google because, um, they can be quite heavy dry reading.

They're structured like scientific papers, because that's how patents have to be structured in the US. There is a very specific formula in a very specific way, and there's lots of scientific diagrams and all this kind of stuff. Even outside of all the usual SEO jargony kind of stuff that you kind of can make sense of, there's a lot of extra, for want of a better phrase, scientific nomenclature that they use. Like, okay, and this is dry as hell. And the way I used to consume these things is the late Bill Slawski doing a brilliant job of just kind of breaking down stuff. Funnily enough, he covered the pattern we were about to talk about back in 2012 when it was originally discussed and released and things like that. And that had always been my way of kind of getting a human to filter that stuff. Funnily enough, I was talking about LLMs and getting humans to filter. I feel like Bill was very much that for me when I was first learning about this kind of stuff. And we're now going around to looking at, as I mentioned at the top of the episode, how to catch spammers out with this really clever little trick called rank transition function. Doo doo doo. It's got its own little theme tune.

Yeah, that was it. Did you not read that bit?

Sorry.

Yeah. NotebookLM might have added that in.

I can add a sound effect.

So this So NotebookLM will summarize the patent for us and give you a little discussion about it. And then based on that is whether I choose to get into the Sahara desert of dryness to actually read it, which I did with this one because it was, I found it quite, uh, as far as patents go, quite exciting. Um, just in that it's a bit wild. So it's, like you say, it's basically, um, tricks that Google plays to try and catch spammers out. And there's 2 main mechanisms they use, which is this rank transition function, which you mentioned, which as far as I understand it is essentially when they detect changes have been made to your site, they will quite quickly be able to assign you a new target rank. So say you're 13th and they think, actually, we want to rank you 2nd. They don't just immediately put you up to 2nd.

Yeah.

They will artificially slow that process down to get there. I assume partially at least to make it harder to directly correlate, huh, I changed these exact 14 words and I moved up 2 spaces 24 hours later, and then removing them has the opposite effect. Um, so that felt kind of expected. To me, I don't know about you. Yeah, I felt like that was a— it just felt like a sensible program-y type thing to do.

You couldn't just, like you said, if you change one thing and then see a direct result of that, that would kind of mitigate basically every SEO tool that tracks all that kind of stuff, right? Like you would just be able to do your own testing in real time, assuming you're getting discovered, crawled, indexed very quickly, of course.

But But you could correlate that against— because you could look at, okay, well, you could direct— came back here and then we saw this after this change. You could create what that window was.

Yeah. So this, that's the rank transition side of things, right? And this is where it comes into the spam catching out people with spammy tactics side of things, which I found very funny and I love the way it kind of frames it. So if from my understanding, if Google catches you doing something dodgy, and I think they listed— I've got a little list here of the things that are in the patent. Keyword stuffing, invisible text, so cloaking, tiny, purposefully tiny text, page redirects that shouldn't be there, meta tag stuffing, and link-based manipulation were the 6 things highlighted by Google. If you do any of that stuff and you get caught doing it, they might purposefully twiddle with your rankings either positively or negatively. And then see if you do anything about that. They're like, oh, you've got 500 links in 2 days. That's weird. They're like, we're just going to fiddle with your rankings. Like, oh, oh, I bet it must have been those dodgy links I built. I better disavow all of those domains. Interesting. So you knew it was spammy, did you? And Google like catches you out and then lowers the ranking and transitions the ranking from that.

So NotebookLM summarized this as like a poker player bluffing. And I was like, actually, that's a really, really good way to do it because Google— so the inverse response transition was Google initially suddenly dropping your ranking before they have the intention to increase it. So it's the same scenario of, okay, you're 13th, we're going to put you 2nd. But tell you what, first I'm going to put you 45th just to see if you blink. And if you blink, I might put you back there. But next time we shift something else, we might do that as well.

Yeah.

Because the pattern then mentions looking for patterns of behavior in these windows of transitions.

Yep. Yep.

So rather than a thing happened and Google's like, oh, that seems to be good. And then they do this rank transition and nothing happens. And they're like, okay, fine. You can have this. A bit like your poker player not, not blinking. And then you're like, okay, well, I fold. Like you've obviously got it. I found that— firstly, I found that a bit wild. So I missed this painting completely the first time around. I never saw this until recently. So again, this is old now. This is like over 10 years old.

Yeah, 2012, when Bill covered it. Yeah.

So again, I'm not interested in it in terms of a, oh, this is a new ranking factor. But I'm interested in it that Google's proactively trying to catch out Spammers and has systems to do that apart from just I'm working on on ranking websites, which which makes sense. And it did make me think because we've seen like this behavior over the years when I've you know when I've ranked sites where there is unexplainable rapid movement which seems to then go back to normal or you know and you just kind of get a feel for that's how it goes and you try and communicate that to clients saying like.

So this is what we're going to say to clients. I was like, sorry, inverse rank transition function, you know.

Well, I think so. I genuinely think, right, I genuinely think, um, so, you know, this pattern notwithstanding, if we had a client where things are, you know, we're doing— we know we're doing the right things and things moving up, and then they have a few pages or rankings, you know, just suddenly go to 40th, like, I would have always said to them, chill.

Yeah.

Don't do anything. Wait for the dust to settle.

Don't go and build 500 spammy links just in case.

Because more often than not, things do right themselves.

Yeah.

You know, you look and make sure—

The dust will settle and—

Yeah, you check there isn't some technical issue like, okay, did we lose backlinks or did the site go down or something like that? You know, or is there a new page ranking in its place? Is there ranking updates? That kind of thing. And sometimes there isn't. And you— I, I've always told clients like, we'll just wait and we'll see what happens, because yeah, 99 times out of 100 it will correct itself. Um, and I think it's helpful to know because you can demonstrably say to a client, well, there literally are documented functions Google has painted that do this exact thing to try and catch people out. So the optimal thing to do is be chill.

Yeah, yeah.

Um, which is cool because not You know, not many, uh, there's not many situations in life where like something bad has happened and the optimal thing to do is nothing.

And there is an actual legal document that is telling you to chill.

Um, and it also, it, it made me think actually a little bit about, because I'm sure that like, I'm sure it's different how it works now, but in terms of like in general the behavior we would see from a search engine in terms of this damped transition. We know RankBrain has a memory, and Google has given us statements about if you improve the quality of your site, it may take many months to see the benefit of that, which to me is essentially something like RankBrain continually getting positive signals. Yeah, for the full time of its memory.

We've, we've seen this before with clients many times. Like, you see that the classic hockey stick style of like you plug away and plug away and plug away and then you get that kind of buildup of momentum, right? Exactly.

Yeah.

Yeah.

It does make me think though about A/B testing with SEO. So if we know there is some kind of dampening of transition, whether it be because of modern machine learning stuff, you know, memory of user interactions over time, or whether it's older school patents like this. How long do we need to run A/B tests for?

Yeah, before— what is that transition window like?

Before we can determine that it has or has not had an effect. And there's probably— and I haven't seen anyone write about this, but I am— my feeling would be there's probably categories of changes you can do that would not necessarily be affected by this. So I'm thinking, for instance, if you change internal linking, which is obviously a super popular thing to do, generally that sees results quite quickly in my experience because Google's less— has to be less worried about internal links because you're playing with a finite set of toys, right?

Yeah.

It's not like external links where You know, you can bring down the power of the sun on Google if you pay. You know, if I had $100 billion and was just paying everyone, I was like, you know, here's your $10 million for a link and here's $10 million never to tell anyone it was paid for, you know, and orchestrated that, you could manipulate search, you know, on a massive scale.

Like billionaires do with Diablo characters, for example.

Exactly. Um, yeah, he got caught out.

He certainly did.

Um, I've totally lost my train of thought. Sorry, sorry.

Um, yeah, you're totally right. The A/B to like chatting to the guys, Nick and Ryan at SEO Testing and stuff like that. And I know some of the studies from Will Critchlow at SearchPilot as well. From what I kind of get is that there's that 90-day kind of— that's the minimum period you should be testing for. Because I think a lot of people, like you said, you want to see fast results. You want to see something shift in a week or a month or whatever it is. And like you said, we always have that conversation of, oh, it can take 3, 6, 9, 12 months to change or to see any improvement. So that 90-day, 3-month-ish kind of ballpark, I think, is a good starting figure. Obviously, the general thing is the more data you can get, the better, the longer you can leave it, the better. But if you start to get— I'd be interested to see if we could, you know, find some examples, either SEO testing, Search Pilot, other A/B testing systems where you get that kind of initial shift and then it completely reverses. Because I've seen a lot of examples with like the Search Pilot case studies. I know we talked about one last month.

Yeah.

you'll see a significant positive impact or an overall neutral impact and all this kind of stuff. I'd be interested to see the graph that Google has on the patent is the dip down, then up kind of thing where it goes below the previous and then up and back to where it should have been, for want of a better phrase. I've not seen any A/B testing graphs like that, so I don't know if anyone has actually experienced this in the midst of an A/B test specifically. I'd be kind of intrigued to do like a canary site like we had a few years ago. It's like, I wonder, do some spammy stuff and see what happens and see if this still works and then A/B test the spammy stuff and see if suddenly changing your mind and reverse it, Google's like, hold on a minute, what are you up to?

I think—

I mean, I don't know what the result would be like. Yeah, the thing that was true 12 years ago is still true.

I mean, Google's spoken as well. saying this year and ongoing, and this was at the SEO FOMO meetup with Gary Ish, that Google will likely be doing more regular kind of updates as well, like algorithmic updates.

Yes. So it's less of the like 3 big ones a year. It's going to be kind of a more smaller rolling continuous kind of thing, right?

Which again makes me think about how do you do you know, A/B testing effectively during these periods. So obviously you have controls when you're doing A/B tests. So if, if something happens during an update, you still have your, I've done nothing to this control. But my point is, what is effective can change during that.

Yep.

So something that was previously effective that you've deployed site-wide might have been because there was a you know, certain edge case in an algorithm which is then updated. So I think that in general is going to become a trickier area, much like a lot of SEO. Yeah, it's become a bit more blurred, which again I think is probably an ROI-positive thing for Google to— situation for them to be in, lack of clarity.

Which leads us all the way back to the Google team's cryptic posts on social media, right?

Don't be evil.

Don't be evil. Exactly. Or don't be evil. Anyway, that wraps up this week's episode of Search with Candid. Thank you so much for joining us wherever you're watching or listening. Go and check out Mark. If you listen to the podcast, you probably already follow Mark, but go and follow Mark on LinkedIn. Go and check out our newsletter at coreupdates.com. That is where Mark and I basically summarize all of the week's interesting SEO news in nice little succinct bite-sized chunks. Also, episodes of this podcast delivered to your inbox every single week. And you're now solicited because they're in your inbox, SEO tips from, from the week all through the whole kind of core updates thing. It's a nice little succinct little sections and chunks that I think we've curated quite nicely. So it's easy to— and we've had compliments on people saying, oh, it's so much easier to read than, you know, a wall of text, all this kind of stuff. Nice little, you know, 3-line headline. chunks of, of here's what matters, here's the gist, go and check it out if you want to read more.

We mentioned this. So the inspiration for Core Updates was Sergio's TL;DR marketing because I used to love that newsletter.

I emailed Sergio earlier today telling him.

Yeah, it was just bullet points and it was just the stuff that was actually kind of important because there's a lot of noise like, yeah, you know, there's lots of interesting things happening in SEO, but when you're working in it, you're kind of like Okay, it's Monday. What do I actually need to do? I need to know to do my job better. And I think that's where there was that gap that you do such a good job as well now finding me the articles where we can say, okay, that's probably worth knowing.

Yeah, definitely, definitely. Go and check out Core Updates, subscribe and like and all that kind of stuff. You can help us out on YouTube and Spotify and all that kind of thing. And if you're going to be at Brighton SEO, keep an eye out for these hoodies. Yeah, because a bunch of us—

Jack will buy you a drink.

Rocket App will not buy you a drink. Absolutely not buy you a drink. Um, but I know you're speaking at Brighton SEO. I'm going to be moderating some sessions as well, which is cool.

Doing your podcast?

Uh, yeah, we're doing a live podcast. So I have mentioned that we did an episode a couple of weeks ago, uh, and we're going to be talking all about building personal brands. And that's me teaming up once again with Sarah and Tazmin from the SEO Mindset podcast. And that's on the Wednesday night, before the main conference starts on the Thursday, uh, in Brighton. Tickets are free, there's free pizza, free drinks, and it's an overall good vibe. Usually sort of like 70 or 80 people there, which is very, very cool. And it's some amazing, cool, interesting— lots of Brighton speakers show up and I'm like, oh my God, I saw you talk last year. That's so cool. Um, so yeah, and we've got plans to do other live podcasts with other conferences and stuff in the future, which is very cool. And I've been planning some potential content to film in Brighton as well, which should be a lot of fun.

So we're doing a couple of things.

Yeah. Remember like when we were last here with Arnoud and Jama and Birgit, and I was like, big plans for 2025. I'm like, yeah, we also— Mark and I have also been discussing an Also Asked podcast as well. It's a whole other thing. So stay tuned for lots of cool ideas and keeping me very busy. Oh, and Search and Orange XL as well. We're doing a conference this year. No big deal.

No big deal. More on that later.

Stay tuned for all of that. Exactly. But thank you. Thank you once again for watching or listening, wherever you're watching or listening, and we'll see you next week.