Navigating Google's Algorithm

Marie Haynes, Brian Dean

Growth Summit 2022

Learn how top SEO minds think about working with Google's ever-changing search algorithms.



[00:00:00] Aadil Razvi: Awesome. Okay. She has been obsessed with understanding how Google determines and rewards quality and websites since 2008. She teaches on the Quality Raider's guidelines, Link Quality, Google penalties, and Google's changing algorithms. Please give a warm welcome to Dr. Marie Haynes.

Thank you.

And we have the founder of Backlinko and the CEO of Exploding Topics, the top resource to uncover high growth keywords before they start trending. His weekly email newsletter is read by over 173,000 SEO professionals and marketers. Show some love for our very own Brian Dean. I wish I had one of those air horns, right?

[00:00:51] Brian Dean: [laughs] Yeah.

[00:00:52] Aadil Razvi: All right. The way that this session's gonna go, I'm gonna start with some high level questions. We're talking about navigating Google's algorithm today. So we'll start with some high level questions. We're gonna get into some more tactical stuff, and then we will cover some audience questions at the end. Does that sound good?

[00:01:09] Dr. Marie Haynes: Sounds great.

[00:01:10] Brian Dean: Let's do it.

[00:01:11] Dr. Marie Haynes: Let's do it.

[00:01:11] Aadil Razvi: I love it. All right. We can't start a navigating Google's algorithm chat without talking about this new Google helpful content update. How should we thinking, should we be thinking about it? Is it helpful? Marie, why don't you take us away?

[00:01:29] Dr. Marie Haynes: Sure. So Google made a big deal about this update. They reached out to several SEOs, including myself before the update happened, to say this is gonna be a big deal, [laughs]. And then when it happened, it seemed like very little actually happened. Now, since then, we've seen a number of sites that have been hit, have been negatively affected by this update. But the site's effective, the scale of the update has not been as large as most SEOs expected. I'm sure we'll get into this, but I think we have a lot more to come to see with the helpful content update. I'm curious, Brian, whether you've seen sites strongly affected. I've seen just a few.

[00:02:11] Brian Dean: Yeah. I haven't seen much. Like you said, Marie, it was a lot of sizzle, not a lot of stake with this update. It was, in my opinion, of PR more than anything, which I think the named updates have been. It's a pattern of Panda and Penguin to improve the results, search results, but also to send a message that we're not messing around anymore when it comes to either with Panda, mass produced sort of low quality, thin content with Penguin, with shady link building. And now, I think with AI produced content was the main sort of message they wanted to get across, that they're not for it. And that... What really confirm that for me was that the update had a slogan like for humans by humans. And for me that was like, it's weird to have an update, have a slogan and a logo and all this stuff,

Which suggested to me was gonna be more of a PR thing than an actual update. But even I was surprised at how little it seemed to impact the search results. It was basically nothing. They announced it's starting to roll out, it rolled out, and it's done. And people were like, "What's that?" But like you said, Marie, there are aftershocks that might be happening. But then, the difficult part then is that it's hard to tie it directly to this update because there's other updates happening all the time. So to me, it was more of a PR thing. That's why they reached out to these SEOs beforehand and warned them, note that they didn't do that with some... A lot of previous updates. They just rolled them out after the fact announced it.

So not to be a conspiracy theorist or anything, but to me it's a real update. I'm sure they rolled something out that attempted to improve the search results, but I think it was more of a warning that they're worried about how AI content is being produced at obviously mass scale.

[00:03:59] Dr. Marie Haynes: Yeah. I think I can jump in there too, and just add that the way that Google described it, they said the helpful content update is introducing a new classifier. It's almost like a new way that they can identify low quality content. And they hinted that it will get stronger as time goes by. So I really think that Google introduced a new way to assess quality. And initially, the sites that were affected were very large AI producing sites that had content that, maybe would do well in old search engines. But when a human reads it it's very obvious to us that it's not human-written. And then when we had, shortly after... So the helpful content update finished rolling out, I believe it was September 9th, it was a Friday.

And then on the Monday, Google launched the core update. And I believe... I don't... I can't say this for sure, but I think that a lot of the sites that had the classifier that are saying, "Okay, here's some unhelpful content on this site," they weren't actually impacted until the date of the core update. So I think all of these updates are intertwined in just more of Google's way of determining which content actually is high quality.

[00:05:17] Aadil Razvi: Okay. So interesting. So you're saying that actually the helpful content update is not really all that helpful. It's more of a signal that, "Hey, this is a direction that we want to move towards," where things like AI written content will be more penalized, or is it less about who wrote the content and more the authority and the topicality of the content itself?

[00:05:47] Dr. Marie Haynes: Yeah. It's less about who wrote it more about how helpful. That sounds obvious to say, but how helpful the content actually is. So one of the sites that I'm soon about to jump into an investigation on is a site that is a massive aggregator of a particular type of product. And much of their content is auto-generated. And that's part of what, the helpful content update goes after. But it's really important to note that Google said it's a machine learned-algorithm, which they haven't said that a lot in the past. They haven't admitted to to using machine-learning to actually determine how to wait an algorithm.

And they told us that it's going to learn as they go along. We're gonna have more of these updates. And I can't see that Google did this and made a big deal, such a big deal about it for it to be just nothing. I think we're gonna see that it gets stronger and stronger as the months go by.

[00:06:51] Aadil Razvi: Is that what you're saying as well, Brian?

[00:06:53] Brian Dean: Yeah. It's hard. I'm seeing what Marie... I 100% agree with Marie. It's just hard to predict the, how much it's gonna impact things. And even if there is a big shuffle, unless Google comes out and says, "That was helpful content update 2.0, to do with Penguin," it's gonna be difficult to tie it back. That's a tricky part of Marie's job. And it's amazing the worksheet she does because it's really hard to... When a site drops, I get this question a lot all the time, "My site dropped. Why?" Even if I had access to Google Analytics, Google Search Console, everything, even then, it's really hard to determine what happens. But if you're just from the outside looking in, it's really difficult to... It's almost impossible to figure out whether it was something on your site, whether it was an update, and even then to tie to a specific update.

And at the end of the day, unless your site is doing sort of a niche thing, I'm not calling it bad. If you're using like your aggregator site that Marie's client, I'm sure it's not shady. They're not trying to do so- it's just the way the site is designed, and I'm sure when users land on the page, it actually is helpful in a way. So it's not necessarily that aggregators are bad or a content is even bad inherently, it's just that they're trying to always improve the algorithm. But usually, they do it a little bit low key, and they usually don't... They haven't done named updates. This is why I'm really honing in on that. It... For me, that's a big deal. Like-

They had Penguin, they had... They had Panda, they had Penguin, they had, before that they had [inaudible 00:09:56] and all these other updates. Then they had the product reviews update. But all the updates since then have either been like generic core updates, or we name them like Medic or things like that. So the fact that they had a whole campaign around this [inaudible 00:10:08], Delta, SEOs it spoke to me as it's more about sending a message than the update itself. Although the update could be significant over time, like Marie said. How significant, it's impossible to say. I bet Google engineer couldn't even tell you because it's machine-learning. So it depends how good it is at detecting this stuff, because a lot of AI content is, especially if it's edited well, it's a little difficult to discern whether a human wrote or not. It still might be low quality. And it usually is, but it's harder to determine whether a human wrote it. And the fact that it was all about human writing, humans by humans, for humans, makes me think they don't want that.

And I can understand why, because air content is the worst of the regurgitation of what already happens. And the search results so much where you have a freelance writer, and you assign them something, and they just have five tabs open on their browser and they just take what's in, they regurgitate it, remix it, and then spit it out on the page. That's already bad. But this is doing it at scale, and there's no editorial process at all. It's just understanding what's already out there. So I can understand why Google would wanna nip this in the butt, and I think that's what they're doing, because AI content's still relatively new. And I think they want to send a message that, "Look, if you're trying to do this stuff, it's not gonna work." Whether or not they'll be able to ever figure it out, I don't know. And it looked like they did, they weren't able to on the first go, at least. But it's definitely something to watch.

[00:09:57] Aadil Razvi: One of the things that Bernard Huang had shared in an earlier session today was that he sees the role of AI tools being actually advantageous in areas where things that people agree on where it's effectively, like facts and more like descriptive type of thing, because it's just remixing what's already there. Versus, it struggles with things like unique perspectives. Would you agree with that? Would you add some nuance to it, Brian?

[00:10:26] Brian Dean: Oh, I agree 100%. 'Cause that- that's all it can do. AI at this point can't create any- It can create new things even Dolly. Yes, it's creating new things, but it's closely based on millions of things that are out there. In a way, not to get philosophical, in a way that's not that different than how humans create stuff. We're all just taking what's out there, remixing it, but there's just a little more direct, and there isn't that spark how to put your finger out. But there's a certain spark when a human creates something, that's not there with AI, especially with AI writing, AI writing is the worst of that... I said it's just literally regurgitating what already exists. It's not adding any unique perspective because the AI doesn't have any perspective, and they've never experienced anything.

So when you're reading something like one of Marie's case studies where she breaks down a site and how it got hit by an update, and how she helped them recover, it's not just about what happened. She's also drawing her years of experience of looking at this. This reminds me of another site that happened, this happened. And AI can't do that. So that's the missing link that AI will always lack, is that real-world experience. It doesn't have any experiences to draw on. And it's not... I'm not talking about fiction writing or writing a novel. Even when you're writing about technical stuff, coding or whatever, you can just tell right away if someone really knows what they're talking about, they've experienced it, or if they're regurgitating. And with AI, it's 10 times easier to notice that.

[00:11:46] Dr. Marie Haynes: It's fascinating... Oh, sorry. To interrupt.

[00:11:49] Aadil Razvi: No, go for it.

[00:11:50] Dr. Marie Haynes: It's fascinating to see how the questions that Google gives us. Brian mentioned, the Panda questions, which they gave us back in 2011, a long time ago. And the core update questions, the product review questions, they all talk about firsthand expertise being important. And that's something that I've really seen with a lot of the recent updates. Some of the recent product reviews updates, you can clearly see that the pages that were demoted in these updates were ones where maybe they, the site owners were trying to demonstrate their expertise. I see people saying "Oh, I've been gardening for 14 years," but there's no external evidence of firsthand expertise.

And so I think that Google really wants to promote people who truly know what they're talking about, and, we keep harping about AI-written content. The helpful content update wasn't completely about AI. I think a lot of AI content is what is initially hit by it. But that's what's lacking with a lot of AI content is they can, it can write the facts very well about a topic. But when it comes to adding that element of firsthand expertise that's always going to be necessary in order to rank for important queries.

[00:13:06] Aadil Razvi: Got that.

[00:13:06] Brian Dean: Yeah. And even for... Oh, sorry, Aadil.

[00:13:09] Aadil Razvi: Oh, no worries. Last word on the topic. Go for it, Brian.

[00:13:11] Brian Dean: Yeah, last word one. With my new company exploring topics, it's interesting because we don't write about... It's all about trends and data and statistics and stuff. So there really isn't a lot of unique perspective. We try to actually shy away from that in a way. There's no firsthand there's no... We don't use any new... We have a policy. We don't even use pronouns because we want it to be very objective. So there's no, they, you, which is usually in sort of content writing, you want to be informal. This is a little bit more like McKenzie style. But the interesting thing is, I'm always thinking about how do I make it unique, have a unique perspective without that person's perspective? Because we have a team of writers.

But what we've really honed in on is that another thing AI isn't good at is aggregating like data and synthesizing stuff. They can regurgitate. But there's a big difference between that and synthesizing. Synthesizing is "Okay, what are the biggest trends in banking right now?" You have to analyze the landscape, see what's there, and add your own perspective, because there's always gonna be some editorialization there. What the AI will do will just take seven, some other articles that are about seven banking trends and just put them in a washing machine and take them out. What we're doing with the writers is they're market researchers who can understand the space at a high level and write about them with all backed by data. So there's tons of data in facts. AI could do that second part, but they couldn't really create new ideas on top of what's happening or identify a trend that hasn't been talked about a lot already.

And that's, like Marie said, that's what Google wants. They want the unique perspective. So if writing technical about a technical topic, it doesn't have to be like, "I did this thing," because we've written tons of posts and our traffics going is growing. You just have to have a unique perspective of some kind. And I think that's... Google is somehow, Marie, able to suss that out. Like you said, with the proper reviews up, they can tell if you tried that spade that you're writing a review on or you didn't. And sometimes... And as a human, it's usually pretty obvious, but it's amazing they can do that with an algorithm.

[00:15:15] Aadil Razvi: Can you tell me like a specific example of where you've been able to elevate the content with that? What may be more of a factual perspective, but adding in that unique lens? What does that really look like in practice?

[00:15:34] Brian Dean: In practice-

... it's basic-

[00:15:35] Dr. Marie Haynes: Go ahead.

[00:15:35] Brian Dean: Oh, sorry, Marie.

[00:15:36] Dr. Marie Haynes: No, you go.

[00:15:37] Brian Dean: I was gonna say, in practice, it's like taking disparate data points and saying, "Oh, maybe this is actually... These are connected, and this is a trend. So for example, we have a lot of food and beverage trends. We have a lot of posts on that and supplements and health and things like that. And there's a lot of interesting trends that they tie together that aren't necessarily obvious. For example, people want like plant-based supplements, and they also want plant-based meat, and they want plant-based everything. So that's a trend that would tie these things together. So if you're looking at like high level food and beverage trends, you have to take a step back and see the landscape and say, "Okay, it's not just plant-based diets people want. People are infusing that in everything."

Or another example is like THC. THC is getting infused in food and beverage stuff. That's something that AI wouldn't be able to identify. They could tell you THC stands for this. It helps with this. It's bad because of this. It's legal in whatever states. They could just riddle off facts. But what they couldn't do is say, a big trend right now is THC infused beverages because this stat shows it's growing. This sta- this data point shows it's growing, and we aggregate those together, and a human is able to make those sort of connections.

[00:16:57] Dr. Marie Haynes: I think there are two ways that Google can assess whether something is written with firsthand expertise. So one, we talk all the time about this concept of EAT, expertise, authoritativeness and trustworthiness, and it's so much more than just a score. And so Google has all sorts of metrics that they can look at to determine does this person actually, or the person or the brand actually have the authority, the expertise to write on this topic? And certain topics are going to need more authority than others. So if I am querying a new medication that I'm put on it's very important that comes from an authoritative source. And so EAT not necessarily having a doctor write it, but the website being known as a trusted authority on these topics is one element of expertise.

The other thing is that I, and I can't prove this, but I do believe that Google's algorithms can determine when content is expert-written just by the types of content that are there. And one example would be a client that we had that's a doctor in a certain one particular area. And he competes against the big players like Healthline and WebMD for his queries. And when he writes about his topics, he's adding things that he sees in his practice, which if you took this to a content writer, they can aggregate, everything you need to know about this medical condition, and it can be the most accurate piece of content out there. But with the doctor adding here's what I've seen, and when we go on this medication, here's what my patients tend to see," it's that type of information that I really believe that Google's algorithms are starting to recognize, "Yes, this is closely associated to this topic, and it's relevant to what users wanna see for it."

And there will be vague things that we can't even measure where Google recognizes when a user needs that level of expertise. So it's very complicated. I see people saying my content is getting beaten by somebody who doesn't even have author bios," when maybe that's not what a searcher would expect for that type of content. And so I think a lot of the changes that we've seen recently have been Google better understanding the intent of the searcher, and what level of authenticity and expertise is needed in the content.

[00:19:19] Brian Dean: Yeah, that's a really good point, Marie.

[00:19:20] Aadil Razvi: Absolutely.

[00:19:20] Brian Dean: I want I... The author bio thing is, I'm so glad you mentioned that, because people think that some sort of magical like thing like, "Oh, that's what I was missing an author bio." And the author bio, I... And especially now in our space, it's impossible to really tell the impact of having a doctor write it, a doctor review it, because everyone's doing it. So there's no control group anymore. And even the site like that got hit with the Medic update, Doctor acts. Okay, maybe he... I don't know his... a lot about the guy, but he was a doctor, and they got destroyed. So he wrote... he was a doctor writing stuff. He's... I don't... Like I said, I don't know if he's good about it otherwise, but he got hit with updates. So it's not as simple as having that.

Plus, you could just make it up okay, I have... Elon Musk wrote an article for exploring topic. I could just put that as an author bio. It doesn't mean anything. It's something that is easy to just put there, but I don't think it sends a message to Google as much like Marie said, it's more the searcher would wanna see that to say "Oh, this is written or reviewed by a doctor, dietician, vet, whatever." And Healthline does a great job at that because every piece is reviewed and obviously very well referenced. And that's one of the reasons they just absolutely dominate the search results in the health space. So yeah, I just, when it comes to EAT, I think an underrated part of it is authority with links to the site because that's another huge factor that Google continues to rely on just because it's so good. 'Cause anyone can do an author bio, but it's really hard to get a link from Yale, Tufts, Harvard, Brown at... It's really hard to do that.

You have to be, you have to be a real authority to do it. So I can see why Google would incorporate that into their EAT, as opposed to just surface level stuff that the quality... A big reason that this thing gets big is 'cause those quality radar guidelines, but those guide, those radars are just looking at the surface level stuff. They don't have access to the algorithm, they don't have access to the links points of the page. They just do some cursory research and it's basically to make sure that they didn't do something really bad. That algorithms totally off in my opinion. So yeah, the author bio, if you're in the health space, obviously is not gonna hurt you, but it's really not gonna make a big difference. You have a doctor review your content as opposed to not in my opinion.

If you have a doctor really review it, then the content itself will be better and then that will help your rankings. But the bio I... Doesn't really make a difference, in my opinion.

[00:21:44] Dr. Marie Haynes: I should add here that-

[00:21:45] Aadil Razvi: [inaudible 00:23:20].

[00:21:45] Dr. Marie Haynes: ... yeah, that very quickly-

[00:21:47] Aadil Razvi: The last word on the topic.

[00:21:48] Dr. Marie Haynes: Yeah. We have had clients that have improved that we think the improvement was because of adding author bios. But it wasn't the author bio that did it, it was the fact that they actually got doctors in to-

[00:21:59] Brian Dean: Yeah. [laughs]

[00:21:59] Dr. Marie Haynes: ... legitimately review the content and provide-

[00:22:01] Aadil Razvi: Yeah, that makes sense.

[00:22:02] Dr. Marie Haynes: ... references and whatnot. So yeah, so I wouldn't discount author bios but I think a lot of people think that's all EAT is when it's just a component of making your page more trustworthy.

[00:22:15] Aadil Razvi: Let me ask you Marie we're talking about how search intent is mattering more and more, the type of result that best matches your query maybe a shorter response matches a query and will be ranked higher in a certain situation versus, long form editorial content. Yeah, just zooming out a little bit where is search going? Google made a big PR move with this helpful content update, but what direction is search moving towards in the next five years?

[00:22:46] Dr. Marie Haynes: I harp too much on this, but Google's blog post on core updates is something that's very important. And that's the one where they list 20 questions. It's almost the same as the Panda questions. And when we first had those questions we would say it's wishful thinking, but like, how could Google write an algorithm to actually measure these things that are in the questions?" When Google di- wrote... When they gave us the Panda announcement they actually said, "These questions give you some insight into how we write algorithms to rank websites." And recently, I was speaking with Alan Kent from Google, who works in the product reviews algorithm department. And I was surprised that he actually confirmed this to me that Google uses machine learning to try to find which content matches those questions.

We don't know exactly how it's used but the idea is that Google can look at those questions is it firsthand expertise? Would you give your credit card to this site? That, those types of questions. And I thoroughly believe that the information from the quality readers now is used to provide Google with good and bad results. Here's a page that according to the quality rate's guidelines would be considered high quality. Here's one that would be low quality. And then Google can use those in their machine learning algorithms to figure out what the weights are to produce the ranking system. So what I'm saying by this is that I think that we're seeing a shift page rank and links as we know it to prop up rankings is still important.

But I feel like as Google gets more semantic capabilities, more capabilities to understand language, to understand which factors are important to users, that they're going to rely more on those and less on links as a ranking factor. Now, links, like Brian said, are still really important, especially from authoritative websites. But I've seen a lot of cases where the types of links that used to work years ago really don't have much impact in today's algorithms. So I think we need to really be focusing on whether our content does a good job at meeting the quality questions that Google gives us. And a little bit less emphasis on link building in a lot of cases.

[00:25:08] Brian Dean: Yeah. I'm not-

[00:25:08] Aadil Razvi: Great. Brian?

[00:25:09] Brian Dean: First disagreement of the day, Marie.

[00:25:12] Dr. Marie Haynes: Tell me.

[00:25:13] Brian Dean: The reason I-

[00:25:14] Dr. Marie Haynes: I do know that links are important don't get me wrong on that-

[00:25:17] Brian Dean: Of course.

[00:25:17] Dr. Marie Haynes: ... but go ahead, yeah.

[00:25:17] Brian Dean: No, it's not a hard dis, it's just like a slight disagreement. I think that they've from what I've seen from the last call, if assuming the trend continues from Google, my opinion is, they're really worried about inaccurate, dangerous, shady AI written content in the search results. In terms of semantically matching results, I feel like of course they haven't maxed it out. There's always room for improvement, but it's really good. It's... And if you use Google Voice search on a Google app, it's scary good. But their main concern right now they're not gonna take a risk and just ch- and flip flop out and take links from whatever it is, let's say like in importance of an 8 out of 10 to like a 5. What they want, I think what they wanna do is hold, serve now.

They own the whole mar- because if you take a step back, they own the market with search. People love Google, people use Google, it's great. There's not a lot of, the ceiling is, they're pretty much there. I'm sure there's more stuff they could do, but it's really good. Their main criteria now is to make sure that people are seeing accurate, good, helpful info. And over time, it's links have been a great indicator to do that. And we've always... Since I started Backlinko in 2013, I've heard that links are gonna become less and less important in two years, less and less important in three years. It never really happened [laughs], so that was 10 years ago. So I think they'll probably be... Of course, they're not becoming more important, and they're probably not becoming the... they're not gonna be the same level importance, so they're gonna decline a little bit, I agree with Marie in that sense.

But I don't think it's gonna be... I think they'll be more or less the same importance as they are five years from now as they are now. Just because it's a great signal for Google to filter out bad sites identify authority. It's just a great signal. And the other stuff is messier in terms of user experience and things like that because one example I always give is say you're looking for way, natural ways to cure cancer. And you go on a, you land on, there's a page at ranks number one. You click on it, and it says all you need to do is stand in your head five times a day, and you'll be cured. Peop- the time on page bounce rate would be really low on that page 'cause people are gonna be reading it trying, they're gonna think this is great.

But it's obviously not true. So Google's real... They can't use those user experience signals that much because people will just believe what- what's on Google. They're not necessarily gonna go back to search and this is really bad. So they have to always def- I'm sure they wanna get away from links. I'm sure they've been trying for years to get out of links, but it's like that, that relationship you can't get out of. They're always going back to them because it's just a great signal to filter the week from the CHF. In my opinion, Google's main priority now is not really like search and 10 or semantically understanding your query. I think they nailed that. It's like pretty, pretty much 100% good at that. It's really making sure really trusted authoritative sites are ranked on the first page, and links are probably the best way to do that still. And I can't imagine another signal coming along.

[00:28:27] Aadil Razvi: One of my favorite things about the black box of SEO is that we can have certified experts in the field just severely disagree from one another.

[00:28:38] Dr. Marie Haynes: [laughs].

[00:28:38] Brian Dean: Oh, yeah. That I never tal- that was a sh- that was a showdown. I'm never talking to Marie again.

[00:28:43] Dr. Marie Haynes: [laughs]

[00:28:43] Brian Dean: We're disagreeing. I'm like... We're dis- we're disagreeing on percentages of like 5% of-

[00:28:48] Aadil Razvi: Oh, absolutely. It was-

[00:28:50] Dr. Marie Haynes: Would you agree though, Brian that some types of links that Google used to value, they're not valuing anymore?

[00:28:58] Brian Dean: What's an example?

[00:29:00] Dr. Marie Haynes: The obvious example would be the links that worked before Penguin, low quality directory links, things like that, that you would really move the needle. And then Penguin came along and Google was good at-

[00:29:11] Brian Dean: Yeah.

[00:29:11] Dr. Marie Haynes: ... ignoring those. And then 2021 link spam update in July Google came out and said that, "Hey, if you're guest posting for links, that's against our guidelines."

[00:29:23] Brian Dean: Yeah.

[00:29:23] Dr. Marie Haynes: And, we can argue, maybe guest posting still works. I don't know. But I really feel like, and from the sites that I've analyzed, there were a lot of sites that heavily used that type of you've built your own link, it might be confused as a natural mention that saw big drops. And I don't think they were penalized for those. I really feel like Google shifted their algorithms to, in those areas to maybe not favor those links as much.

[00:29:50] Brian Dean: Yeah, 100%. I actually agree with you 100% there. They- they're just... I think they're really just filtering... Unless you're doing something really egregious, it's more or less filtering out those types of links that you built and then you drop and you're like, "I got hit by an update." But it's really just you're, on stilts and they pull the stilts out from under you. That's more or less what happened. But in a way that makes every link you do build legitimately more valuable because it's, it makes links almost more valuable because it's harder to just build a bunch of stuff. If you get a really good solid link, that's worth way more because the other stuff doesn't even work anymore. So that's why I'm seeing links are still for me, I don't think they're really gonna be that much less important.

Of course, it'd be a little less important in five years 'cause they'll do all sorts of other signals that they're able to refine. But I think we're talking like 5% less important than they are not like 50. Five years is a long time, who knows. But I agree with you, Marie, that anytime you're relying on a type of link, you're probably in trouble in my opinion.

And if you're building links with your desired architects, your desired page that's usually not gonna build well over the long term either. So when I say links, I'm talking about the legitimate links. A little guest posting is not gonna hurt you, I don't think. But yeah, if you're doing it at scale, it will. And, but, yeah, that's my counter that sort of, when you build a good link, it makes th- that much more valuable because other ones don't really work. Like you said, they don't work anymore.

[00:31:12] Aadil Razvi: It's interesting that is quite contrary to what Bernard and Eli were saying earlier today. And Eli even mentioned that he goes head to head with you on this topic. And so yeah I could go down that rabbit hole for the next 30 minutes, but we'll switch gears a little bit and get tactical here. Let's just suppose that an attendee's website completely tanks after a major core update. What's the order of operations? What do they do first? How do they diagnose the issue and plan a path forward? Marie, why don't we start with you?

[00:31:50] Dr. Marie Haynes: Sure. So first thing is just to look at whether you've made any changes yourself. I've seen many cases where it just happened coincidentally, that Google ran an update on the same day that we decided to do a site redesign. And then it can be very difficult to know whether it was the changes you made or what Google changed in their algorithms. And so if you've made significant changes to your site that's where you wanna look. But if you're affected by a Google update, like a core update or product reviews, in most cases, you're going to see that your decline in traffic and rankings happens pretty much the day that the update rolls out. We've had a little bit of confusion where some of these updates, they've had a burst of activity on the last day of the update rolling out.

But pretty much if you see like a massive drop in rankings and traffic that lines up with a day that Google ran an update then you can assume that the update was because of the change in Google's algorithms. Now, the next question is how to fix that? And it's become really hard to... There's no straight answer to that. In my experience, it's rarely a technical issue, if you were impacted by a core update technical issues that impact like Google's ability to crawl or render a page can certainly impact rankings. But with a core update, it's all about content. And again, this blog post that I always harp on about blue Google on what site owners need to know about core updates.

They tell us there's two things we need to focus on. One is content and the other is EAT. And so usually with a core update, we can find, and lately it's been more content, more how well does this page answer the searcher's needs than EAT? But both of those are important.

[00:33:44] Brian Dean: Nothing to add there, Marie's, she knows about this 10 times more than I do.

[00:33:50] Aadil Razvi: Cool. Then Brian, I'll ask you about this new trend that we're seeing around short form video. The TikTok application of the internet as we are seeing now we're seeing TikTok results showing up in the searches, YouTube Shorts are now getting prioritized. Where is this going? How do we think about video's role in SEO over the next 18 months?

[00:34:14] Brian Dean: I can see it being a semi big part of it. Not over 18 months. I would say it's more of a longer term thing.

[00:34:22] Aadil Razvi: Okay.

[00:34:22] Brian Dean: Because you got to... The thing about these short form videos is the vast majority for entertainment. Very few are for practical use. And Google has dipped their toes into this before, not a lot of people remember because they stopped doing it. They do it still a little bit, but they had video featured snippets for a long time. They still do, but they were prominent, more prominent in the past, and I think they discovered it's really hard to match search intent to a video. And what they would basically do is, if you had a vi- if you were like, "Oh, how to peel garlic?" They would take a video about how to peel garlic, but they would take the section that they thought was most relevant to the search and pull that out.

And instead of hitting Play on the video, it will... Instead of starting at zero seconds, it would start at 15, when the person got through their whole spiel about blah, blah, blah, blah, it would get right to the peeling of garlic. More or less, a TikTok length video that they're trying to pull out of YouTube. They never said anything about it. They it came and went. It still exists when you search for certain keywords, but when I saw at first, I was like, "This is gonna be a game-changer because videos can be really good for certain searches, but there's a lot of fluff around it usually." There's a lot of "Hey, I'm Brian Dean, and blah, blah, blah." People just want for certain things, just want and things. So in a way, TikTok style videos are perfect for that because they're short, there's no intro, they're to the point, and sometimes they just jump into the middle of something that's already happening, which is in some ways perfect.

So I... But I don't think it's gonna be that huge, just because first of all, TikTok is, like I said, the video's a 90% entertainment. So even though Google is amazing at building an own algorithm, they're gonna have to build an algorithm on top of theirs just for TikTok and YouTube Shorts to pull out those 0.01% of videos that aren't entertainment and that are useful, and then match those to the query, and then of that group find the best one. Are they really gonna do that? I don't know. They alrea- like you said, they already are incor- testing this sort of out. But I just think in, for most queries, people want texts based content, or they'll just go to YouTube or go to TikTok.

People want texts for the most part in their Google search results, and they've pushed YouTube and they've have YouTube results of course, in a lot of searches, but less than I thought. And if you look at this is another 10 year ago thing. If this, like 10 years ago, video was gonna dominate the search. You just need to... If you didn't have videos and the Yost video SEO plugin on your site, which doesn't even exist, that shows you how anymore, that shows you how it's not important. If you didn't have that you were gonna be irrelevant in the ses. 10 years later, there's still about the same amount of videos as there were, there was then, even though the number of videos has exploded and their ability to understand videos has exploded.

So for me, I think it'll be part of it because like I mentioned, it has utility. If you can find that video that just jumps right into your problem and answers it in 20 seconds, it's perfect. I think it's gonna be tricky to find among the sea of just entertainment videos.

[00:37:33] Aadil Razvi: Good take. Definitely not something you hear often. We're seeing like a lot of talk around how this is the next... There's a lot of buzz around this. And so that, that's a sobering perspective. Marie, how do you feel about it?

[00:37:48] Dr. Marie Haynes: I think we need to pay close attention to video. And the reason why I say it is because last year at Google's Search On event, and that's where they announce, "Here's the big things that we have coming up." They talked a lot about MUM, M-U-M, this natural language processing framework that's more powerful than BERT when BERT first came out. And one of the things that they really focused on was how MUM was really good at identifying entities in videos, even if it wasn't mentioned. And they showed a video that was all this stuff about a penguin and doing all these different things and they said, "They could... MUM could identify that this video was about the life story of a penguin, even though those words weren't mentioned in the video.

And since then, we've seen, like Brian said, Google show a search results where they can jump right to the answer in the video. I think, and then Search On event this year, Google announced all these changes. A lot of them were related to shopping but changes in the search results that I think we're gonna start seeing more people stuck in the search results than going to websites. And one of the things that Google's probably gonna try to do is to get more people on video, especially in regards to product reviews. There's been so much information all these product reviews, updates that Google has coming out. One of the ways that you could thoroughly demonstrate firsthand expertise is by actually videoing yourself reviewing products.

And I think that's something that Google is gonna start to reward more because I see it in the chat here that people are saying, "The younger generation, some of them don't even use Google."

My kids, if they want a recipe, they go straight to TikTok and-

... And they're gonna be adults soon. So Google knows they need to adapt to that and they'll be playing around with it for quite a while, I'm sure.

[00:39:42] Aadil Razvi: Brian rebuttal to this

[00:39:44] Brian Dean: [ laughs] I think we'll just have to see. I'm always skeptical about like, when Google announces at these events, because we've seen tons of things that have announced. Stadia just got like closed yesterday, and how many, they had all conferences just for that? ANP was like the next big thing for mobile. And that's getting to appreciated. Not to... My issue... I agree with Marie, they're gonna wanna do these things. My reluctance is, can they do it? And I wouldn't discount Google 'cause they can do amazing stuff with their technology, don't get me wrong. But I've just seen this sort of... I call it hype cycle, but and I make videos. I know the space. I think there's potential in it on the platform you're on. I just... They haven't done a good job of incorporating that into the search results.

And short form may change the game because it is... It's perfect in a way for searches because it's a short answer, and it jumps right in as opposed to a long YouTube video. It's easier for them. The other challenge is, ho- I think the... If they're gonna do short form, it's gonna be less of TikTok and more YouTube Shorts because they're not gonna build a whole algorithm around another platform like TikTok. It just doesn't make any sense. They've been able to have a, they had a pipeline to Twitter that they lost, so they're not gonna make that mistake again, in my opinion. But they have YouTube Shorts. It's still really small compared to TikTok. And then basically TikTok re-posts right now, same with Instagram Reels. So I don't know. My take is, I think it'll be a thing, don't get me wrong.

I'm not saying video's gonna be less important, but I'm always more of the old school. I've just seen this like sort of revolution and the search results ha- so many times. And if you look at the results from the year 2005 to now, they're basically the same. There's some ads in the top, there's a bunch of blue links that you click on, there's some features, and now there's more stuff. Obviously, as featured snippets, like Marie said, if this is a way to get them, keep people on the search results, it'll incentivize them obviously to do it more. And there's more bells and whistles, but it's basically the same. You type in your thing that you want to search for, you get a bunch of results, you pick one, you click on it. If you don't like it, you go back to results. I think in 10 years is pretty much what Google is gonna be.

[00:42:03] Aadil Razvi: We're gonna rapid fire for the remaining next five-ish minutes. So we'll try to keep it punchy. Short form as as Google is trying to learn. First, right off the bat we'll go with something high level here. An audience question. How do you create content that withstand any of these major Google updates? How do you future-proof your content?

[00:42:32] Brian Dean: Go ahead Marie. I think you-

[00:42:34] Dr. Marie Haynes: [laughs] I was gonna let you go first.

[00:42:36] Brian Dean: Okay. I

[00:42:38] Dr. Marie Haynes: mean-

[00:42:38] Brian Dean: Oh, sorry.

[00:42:38] Dr. Marie Haynes: Oh, you go. No, you go.

[00:42:41] Brian Dean: No, please.

[00:42:42] Dr. Marie Haynes: Go. Content creation expert. You're the guy.

[00:42:43] Brian Dean: We could do this for 15 minutes. [laughs] We'll just do the rest of the session. Just a trade off. So the number one thing, I think it's more or less what Marie had really harped on in the beginning. It's that unique perspective. It doesn't completely protect you from an update. There's nothing you can do that's "Oh yeah, I can just sleep at... No, I will never get hit by any update ever." It's always possible, but it's extremely unlikely. If you're writing helpful original content that has unique perspective from experts, if you do that, it's gonna be really hard to get hit by an update. Google won't wanna hit you an update because you are providing what they want. They want really good quality content that raises the bar because that makes their search engine better. It makes people go to Google. So if you give Google what it wants at the end of the day, usually they'll give you what you want, which is traffic.

[00:43:34] Dr. Marie Haynes: That's good. That's good. And I think that, I totally agree with Brian. I think that the question is not possible. I don't think that we can make content that withstands any Google update. And the reason why I'm saying this is, often, Google will understand their... They'll change their understanding of intent. So say you had the best article, informational article on a particular product, and Google decides that most people who are searching for that product just want a transactional page. They just wanna buy the product. You can write the best article ever and you're not going rank in a transactional spot.

In some cases, Google updates can be devastating because they've decided that user intent is different. But I think if I wanted to write content that has the best chance of withstanding Google updates, what I'm gonna do is number one, focus on the user, which sounds like a corny thing to say. But this is a trend that I've seen amongst when we look at pages that lost and one with core updates is that the pages that win are the ones that quickly get the user to their answer.

If you think of how we search, you're on your phone, you're... You don't wanna read a 4,000-word article on this topic, you just searched for. The vast majority of the time, you wanna scroll and go, "Ah, here's what I was looking for." And either read that or move on. Number one is finding that. And then number two is to always be keeping an eye on what it is that Google promoted. So if you notice that you've lost rankings, then start looking at the content that popped up ahead of you. And what I find is hard for people is to look at that without saying, "Oh Google screwed up here." There's always a reason, and it might be a combination of things, but there's always a reason why whatever is ranking better is better for searchers.

And so if you are writing always with your users in mind, then that's your best chance of producing good content as opposed to just chasing what certain tools might say or chasing metrics. I think having the content that actual searchers want is by far the best thing we can do.

[00:45:42] Aadil Razvi: We helped Google's algorithm rank using any sort of Google ads or paid social, or are those signals outside of the organic rankings?

[00:45:56] Dr. Marie Haynes: Paid is-

Google's vehemently said that there's no connection between paid ads and organic search. I know people all the time will say I stopped spending money on AdWords or on ads and and now my organic rankings dropped." Social I think is the same way. However, content that often, content that tends to do well on social is the same type of content that does well on organic because people like it. But I think that any correlations you can find there are more just correlations rather than, it's not that, "Oh, you went viral on Facebook, therefore you're going to be ranking better," unless going viral on Facebook brought attention to get you more, better authoritative links. That's a possibility there.

[00:46:44] Aadil Razvi: Brian?

[00:46:45] Brian Dean: Yeah, I agree 100% with Marie. They're totally separate. And not only has Google said it, I always take what they say with a little bit of a grain of salt, but in this case, it just makes logical sense. It would be a massive... It would be close to a scandal if they, it was discovered that there was some sort of connection between having outwards account, having, using a Google ads and your rankings in search results. Besides, take that outta the equation, which is true. It would make the search results worse. They want to have people go to Google but when... It's same thing with YouTube. A lot of people are like, "Oh, if you run YouTube ads, does it help your-" No, it doesn't help your videos do well in organic because they just want great videos on the platform that you wanna watch 'cause you'll stay there.

If they start adding all this artificial stuff, in the short term, I'm sure to give them a little bit of boost. If Google Ads tomorrow, anyone who's running Google ads ha- get... their rankings move up one spot. They'd probably have record breaking [laughs] amounts of people entering their credit cards into Google Ads. But over the long term, it would hurt them because that- that's not... It doesn't make their results any better. It makes it actually objectively worse. So yeah, there's no connection there. And in terms of paid social. Marie said, the only way that can help you is if it leads to links, and it rarely does because only a few people, if you look at the whole world, can add a link to your site.

The vast majority of people, like my mom, she goes on Facebook, she posts about politics, she texts me, she goes to different, the new, the local news website on her phone. She plays Candy Crush. She doesn't have a website, she has no way to link to me. even if she saw my content, this is the best thing I've ever in my life, She can't link to me. So with paid social, a lot of times you're just getting in front of, normal people who don't have websites who can't link to you or don't work at a website that can link to you. So paid social, in my opinion, is good if you, if that alone is your goal to get branding, to get people to come to your site to get conversions. But in terms of getting links from it, it's so rare. And I've tested it a million times because I'm like, "Oh, you can just pay money and get links." And it's not, it's white hat. That's the goal, that's the holy grail.

But it very rarely works because it's hard to get in front of these people. I call them link creators. Link creators are like people that can create links. There are small segment of the population. And that's why you need, you usually need to do outreach to get links because you identify those link creators and you email them directly and basically ask for a link. There are other ways to get links passed from link creators, but it's not usually from any sort of social media thing because it's hard to target. Those folks.

[00:49:20] Aadil Razvi: Got that. Yeah. The big takeaway in the chat was that your mom plays Candy Crush as popular [laughs] inside.

[00:49:27] Brian Dean: Addicted. She's yeah, she like, still plays it like it's 2010. I think she's-

[00:49:31] Aadil Razvi: [laughs].

[00:49:32] Brian Dean: [laughs].

[00:49:34] Aadil Razvi: I love it. I love it. Next year, she'll be on Floppy Bird.

[00:49:37] Brian Dean: [laughs] Yeah.

[00:49:39] Aadil Razvi: Very cool. Thank you both so much for spending your time with us today. Definitely wanna give you each the floor to tell the people what you have going on and how they can get in touch. Marie certainly share with us a little bit about how people can learn more.

[00:49:54] Dr. Marie Haynes: For sure. First of all, thank you for inviting me to do this. I always love babbling on about what I think Google is doing. If you wanna reach out to me, you can do so at And you can go to my website and find, I have a podcast that I do every two weeks where I babble on about what I think Google is doing, and how we can take advantage of it as site owners. And I also have a book on the Quality Raters guidelines that many websites use or many site owners use as a checklist to assess your website in the eyes of Google. You can find that And one final thing I'll just add, I'm recently starting to add strategy sessions where you can book an hour with me to discuss your website and talk about the latest Google algorithm, updates, and changes, and we'll talk together about how to make improvements

[00:50:49] Aadil Razvi: Insanely valuable. Thank you very much, Marie. Brian?

[00:50:53] Brian Dean: First, I wanna say that thanks for having me. Thanks for chatting with me, Marie. I... anyone who wants, who's serious about, protecting themselves from Google updates and being prepared from Google updates should book that session with Marie because I've been in the game for a long time. Marie's the best. And I can almost guarantee that you'll get your money's worth from that session because you can at least feel protected and know that you're doing whatever you can to prevent yourself from getting hit from an update and probably just improving your site in general, because Marie has a lot of good insights on just how to create a really helpful website.

[00:51:29] Dr. Marie Haynes: Thank you.

[00:51:29] Brian Dean: Yeah. In terms of where to find me at @BrianEDean with an E on Twitter my company Backlinko got bought by Semrush my.. So I'm always putting out courses on the Semrush Academy, so I recommend checking that out. And my startup exploring topics is a great place to discover trends before they take off. So check out and sign up for a newsletter. We have about 100,000 people that get our trends every week on that.

[00:51:57] Aadil Razvi: Fantastic. Thank you both tremendously. As you all can tell, we only have A plus players on stage at the summit this year. And Brian and Marie are no exception. Thank you both very much for your time.

More recordings from Growth Summit