INBOUND 2016: Rand Fishkin "How To Keep Up With Google" + Drive Organic Traffic from Google

Howdy, gang. Welcome. It's our last day of inbound and possibly the last year before my people get taken into camps. So that's. There's nothing but gallows humor this week. I didn't have time to put a bunch of election related jokes in the presentation. I apologize. So just you'll have to use your imaginations and I unfortunately don't have time for an extended introduction because we have so much amazing SEO stuff to get through. All right, so what I want to do, what I want to do is to help all of us keep up with the massive amount of change that is happening in Google's world and in the SEO world. And that means in our world, whether you are producing content, helping to amplify content, helping people get to your website, helping people answer their questions, what marketers do best is help people. And to do that, we need tools like Google to work in our favor and therefore we have to understand. So I've got all these slides online at SlideShare. You can find them in there. Bit lykeepup2017 Bitly keepup 2017. I'll have this URL at the end of the deck too. Okay, so we have seen Google in the last couple of years make an incredible amount of changes. And that includes things like, you know, massive numbers of algorithmic updates, this big move to machine learning first in terms of their approach to engineering. Right. We've seen sort of the end of ten blue links. This. It's like spotting a snow leopard in the wild. Like, you just don't see results like this anymore. It's crazy. In fact, I think Mozcast has this at something under 3%. So about 3% of all of Google's search results that people actually look for, not random, weird long tail queries come up with something like that. AdWords has been taking away some of the most critical data. Many of you, I certainly would go to AdWords and I would look for how often is this term or phrase searched for versus this other one. That data is now much, much less precise, much less accurate. And Google is hiding a bunch of terms. I'll show you that later. We've actually seen PPC growth start to level out. This is true. More on desktop than on mobile, but it's happening on mobile as well. So this is not to say that PPC is going down. Google is making more money from it, and there are more people using it and more clicks. But the growth rate itself is slowing, suggesting that acceleration has sort of plateaued. In paid search, Google keeps Fighting to increase the number of clicks that they get to paid versus organic by making the ads more and more subtle. But you tell me, how do you make it more subtle from here? Seriously, look at those top two and the bottom two. Do you see the only differentiating factor? It's that little green. Like, what are they gonna do next to make it more subtle? Remember when it used to be like, oh, it was clear which ones were the ads and which ones are not. So over time, what Google has found is that people learn and they bias their behavior. So every time they make a change to the ad format, right? And ad visuals click through, rate goes up. And you can see that in their quarterly numbers. And then when people get accustomed to it, it goes down over time. I think Andrew Chen had a beautiful word for this. He calls it the law of shitty click through rates. Right? And that and that law, which I think we can probably enshrine in science's halls of marketing, is basically when something comes out and it's new, it will get a lot more activity than it will over time as people become familiar with it. Let's just hope that Macedonian teenagers on Facebook realize this soon. So thanks to, no, not big news junkies like me. All right, there's one person. I appreciate you, sir or madam. So thanks to clickstream data. This is one of the unique things that Moz has been doing that I think is really cool. So clickstream data is essentially, I get asked this a lot. How do we get it? So we basically buy it from these providers. We can get data about all the pages that people visit on the web, desktop and mobile. Because some applications that you install on your desktop, some applications that you install on your mobile device or browser extensions or those types of things track all your webpage visits. And they don't care whether you're on a secure website or whether you're in a secure browser or whatever. They track it all so they know every URL that gets visited and what path. These clickstream providers essentially anonymize that data, sell it to us in aggregate, and then we analyze it to see what are patterns like. So what happens on average on the web for searchers and users. Which means we can now get data like this, which is super cool. We haven't had this in years, not since the search engines themselves were giving us this data. So we know what percent of people are searching for 1 word versus 2 word, 3 word, 4 word queries, 5 word. This has gotten, by the way, this data hasn't been out, at least not In a format that I've seen and trusted since like 2006. So it's been 10 years, but it's leveled out a ton. It used to be all concentrated in 1, 2, 3. Almost no one was searching for 456. So that's changed. I also made this graph, which I recognize in a room like this, a little hard to see, but it essentially shows us the number of searches that happen for particular keywords and then the percent of the fat head, chunky middle long tail of the demand curve that's taken up by those. So when folks are talking about like, hey, we should play in the long tail. Well, why should we play in the long tail? Well, it turns out the top billion keywords that people search for, like the Most searched for 1 billion keywords in Google only account for 35% of all searches. That's kind of cool. Like, the awesome part about that is what unique, weird snowflakes we all are when we get in front of our search box and start typing. Like, there are a ton of queries that are just almost done very, very seldomly, but they make up for a huge percent. When taken together, the long tail theory may have fallen down. Chris Anderson's long tail theory might have fallen down in music and in other forms of demand, but in search, it is still quite strong. I think that's pretty exciting for us. In terms of search activity, average searcher is doing about three queries per day on desktop. I don't have the mobile numbers for you right now, but the distribution of those clicks, this is important because there's some caveats in here. So about 1.19% of all searches, all searches, not all clicks, but all searches result in a click on a paid ad. So Google is making. What are they making? 45, $46 billion, I think was AdWords approximate revenue or for this year. And that is driven by 1% of all the clicks. Sorry, 1% of all the searches. That's crazy. That's incredible. Like, Google is making all their money on 1% and then serving the rest. You know, the 99% of searches. They're just giving us answers. That's an incredible business model. 51%. This is clicks now. So remember, not searches, but clicks. About half of all clicks, a little less than half of all clicks of all searches don't actually result in a click. Like they're answered before. I'll talk about that in a sec. So this is 51% of all clicks that happen on. This is Google US data that we have here. Those go to organic, non Google results. So when you say, well, what is really the opportunity for SEO in the search world anymore, given how much Google is taking for themselves and how much they're taking for advertising? The answer is about half. About half of all searches lead to an organic result. But 49% of clicks that happen in searches go to something Google owns, one of Google's properties, or a Google Ad, or YouTube, Google Maps, Google support page. And there's also a lot of searches that result in no clicks at all. About 40%. 40% of searches result in no clicks at all. For a query like this. I bet Google has trained you to do a bunch of searches, right? Like if you were looking for sports scores last night for Thursday Night Football, you could easily see that without ever having to click on a result. And Google knows that that's about 40% of all searches with zero clicks whatsoever. The query is answered by the engine and then search suggest, I think is biasing some of this. Right? So some of this behavior is as you're typing. Google's not keeping up. About 25% of desktop queries are coming through Chrome Instant right now. This is a list. Sorry for the return carriage there. This is a list of the top 20 sites and they account for 22.8% of all Google search traffic referrals. It is good to be the king. I don't know what number 18 is, but the rest of them. You really want to be one of the rest of these. By the way, shout out to two folks who I think are very impressive in here. Pretty remarkable to see Netflix, who was nowhere, like Netflix was nowhere. And I think that's big credit to Laura Lepay, who's their. Their head of SEO. Done a great job distribution. By the way, the top 20, I showed you 22.8 as far as like top 100, top thousand, top 10,000. So this is very different than what we saw with the search demand curve. It's not like billions of websites are getting tons of traffic from Google. The top 10,000 sites are getting almost two thirds of all of Google search traffic, which can be a little scary. Okay, so we know there's been a bunch of change. What does that mean for us? Well, let me try and talk through the big five and see what kind of difference those are going to make to us in the years ahead. So it's not just Moz that has clickstream data, right? Like, Google obviously knows all of this. They know everything about where traffic flows thanks to owning Chrome. I think that was Actually why they developed Chrome is because the search engineers said, hey, we need all this data on where everyone goes on the Internet, right? And on Hub and Android, right? And so all of these strategies are yes, revenue lines of their own, fantastic businesses of their own, but they are walls around the castle of search, right? 90%, I think it's maybe 88% this year of Google's revenue is still coming from search. And they need that, they need to protect that. This is one of the ways that it makes it incredibly hard for folks like Microsoft or Yahoo or DuckDuckGo or anyone else to compete. So they know where a visit path begins and they know where people go next and where they end up. Right? And so they can get some sense of when people are satisfied by a query. If Google knows that your site, for example, is getting 80% of their traffic from them, they might kind of go, well, is that a good thing? Is that a right thing? Does that suggest maybe that there's some challenges with your model? And I would urge, folks, if you want to do SEO, if you want to do well in SEO for the long term, I think you have to diversify your traffic sources. I think Google gets pretty suspicious when they see it this way. This is data from similar web who also collects clickstream data. They have a big set of providers to them as well as their own properties like the similar sites, plug in and other stuff. And you can see this is the broad distribution of global, they call it online marketing channels. This is essentially how people got to websites. This is like globally how people got to websites. So direct is still the most popular. Which homepages are a big part of that obviously, right? And auto loading pages. But search is next, then referrals. It's kind of crazy how small social is, right? I mean like I get, you know, the whole Mark Zuckerberg like keep everybody on Facebook, on Facebook, don't let them leave, build the walled garden. But that is quite successful, right? Like when you go to Facebook, you really stay on Facebook. When you go to Google, you leave Google and go somewhere else. Video has been rising pretty fast. Although I am skeptical of Facebook's view numbers. Remember, a Facebook view is counted at three seconds, which I think could just be slow scrolling, whereas a YouTube video is counted at 30 seconds, which that's pretty slow. I would say that those numbers are probably not totally comparable. YouTube, however, is the world's second largest search engine behind only Google itself. Clearly they built something impressive there. DuckDuckGo, however, is our fastest growing search engine. Props to Philadelphia. Way to go. Startups love it. Amazon is probably an underrated search engine. In fact, a ton of searches happen on Amazon. We're looking at that now with the clickstream data. And we can see that a tremendous number of e commerce centric searches happen on Amazon, not Google anymore. And so I think the takeaway here is these things are growing, but they're also growing in diversity. We can't ignore a search channel entirely just because it's not Google. One of the best ways you can go about this, if you're using a tool like SimilarWeb or others. Jumpshot is a popular one. You can go in and see where your competition is getting their traffic. So I pulled up Warby Parker here and I can see they get it from direct from Google search, from Facebook, from. Who's that? ShareAsale. ShareAsale.com must be a nice affiliate relationship there. And, and this is a beautiful way to see whether folks in your field, your industry, could be getting their traffic from other places. If you are going down the path of investigating YouTube ranking factors or Amazon ranking factors, there's a bunch of good lists of those. I won't get into them today, but there's a tremendous amount of content for you to be found. You're going to have to play in Amazon. Specifically, you're going to have to play with conversion rate. Conversion rate is one of the things that Amazon looks for because obviously that tells them a searcher was satisfied. I think one of the things we're going to have to do in years ahead is also consider content for multiple engines. So a few of you probably have seen Whiteboard Friday, I suspect. Whiteboard Friday, one of the weird and interesting things we do with it is we put it up first on Moz using Wistia, local Boston startup, and then three months later, we'll put that same video on YouTube. Why is that? It is because we want to make sure we own the Google search traffic. We get the links, we make sure Moz is the place where, you know, you can view it first. We do not want to build a brand of people going to YouTube to watch our content. But we also know that a tremendous amount of search activity takes place on YouTube and we can't ignore that. And so we put up those videos a few months later as essentially a way to have our cake and eat it too. It's not a perfect solution, but we found it to be the best for us. And you might consider this in whatever forms of content you're creating and promoting. I do exactly the Same thing with Medium, by the way. Medium's a phenomenal distribution platform, but I always want to publish first on my own website because then I don't get access to the cookies and I can't control the experience on Medium and I don't get to retarget and remarket to all those people. And retargeting and remarketing are incredibly powerful marketing tools. All right, number four, we are going to have to evolve some of our keyword targeting. In fact, a lot of our keyword targeting, because Google has become massively more sophisticated about this stuff. So here we go. On the left and on the right I have searched for do I need polarized sunglasses? And Google is giving me the answer. In fact, I kind of don't have to click through, right? They're doing this for a ton of queries. They have, over the last year, changed the desktop experience to mimic the mobile experience. So you remember the multiple column layouts that Google used to do, especially with maps and local stuff, they have really shrunk that into single column format so that it can match. And now they're putting more and more of these answer boxes. Some folks call them instant answers or knowledge results or those kinds of things. Ranking number zero, right? Which is, which is pretty sweet. And sometimes the idea is that you don't have to ever click that third party site. Now, we expected that when this one popped up, you know, title tag, how long should a title tag be? We expected that Moz would lose traffic, right? Because Google's answering the query before you ever have to get there. So, you know, it says right up at the top 50 to 60 characters, what more do you really have to know, right? You've got your answer, you can go on your merry way. Well, we actually got more, we got more traffic by ranking in position zero than just ranking in position one. Well, one and two, I think someone at Moz must be good at SEO. So the point being here, right, that this, this is cool. This is good to know that we can take these boxes. And it doesn't always mean that Google is stealing the traffic from us. Sometimes it does mean more, sometimes it means less. So carousel results are ones where a lot of sites have been massacred. Just their traffic has been decimated by the appearance of these things. We are estimating from our clickstream Data that about 50% of all the searches that used to go to the organic results over here are now going to these knowledge graph results. So that's a huge cut in half for those folks. Organic results on a page like this, they're probably only getting about 60% of the clicks from searchers, but here I know 3% of results, but 100%. And yet when you do your keyword targeting, when you choose which keywords you're going to try and go after. Very, very few of us are thinking about relative click through rates and that is probably a dumb and bad move for our marketing. Right? That's like ignoring the fact that spending the same dollars on an ad, on two ads, knowing that one of them is getting almost half the clicks that the other one is. You would never pay for traffic that way. You shouldn't optimize for traffic in that way either. Right. And so I think we're going to have to start doing some estimating in our keyword research. You could do that yourself or you could use a tool. Moz Explorer is one of the few places that that does this in an automated fashion now. But I'm sure in the months to come Semrush and Aahrefs will also be buying clickstream data and putting these sorts of metrics into their tools so that you can choose smarter. But you can't just use AdWords. That is very clear to us. It's possible, by the way, to use SEO to get into these featured snippets. We've done some very cool experiments and results on Moz and you can read more about them from Dr. Pete, who's done a ton of experimentation on this. If you want to rank number zero, he's there to teach you how. All right, so volume data, that's another one where Google is killing us. Classic source of volume data traditionally has been AdWords. They show you these numbers in your account. Probably all familiar with those. Well, you know what these numbers are? You might have suspected this already. They are not real. Right. They're sort of like Google's results numbers. Like they represent a range of potential outcomes and they're not, I don't want to say they're averages because they're often not in the middle of the range they represent. So they're not medium. They're like somehow randomized. It's very, very odd. We did a bunch of work to analyze how many clicks or how many impressions adwords when you bid said the phrases actually got and then what AdWords showed in the monthly data thereafter and essentially discovered these numbers mean ranges. And the ranges are inconsistent at best. Trends is a little bit better. Google Trends is a little bit better. Tends to be the case that if you pop something into Google Trends, you will See more accurate ish numbers. Unfortunately, without a left access there, you're stuck guessing what the numbers might be, but you might know what's more and less. The other thing Google's doing in AdWords that's horribly frustrating is they are combining what they call similar terms into the same bucket. So they'll say that flowers, San Fran Flower, San Francisco, San Francisco Flower Market, San Fran Flowers and San Francisco Flower Shop, which in my opinion are all different queries for different hopeful outcomes. Like when I search for San Francisco Flower Market and San Francisco Flower Shop. I mean two different things. But Google is grouping these and giving us volume numbers for all of them combined. Which sucks, right? For a marketer like this takes away a lot of your keyword targeting prowess. Other times we're putting in keywords. So I put in. What did I put in here? Fitness tracking. Fitness tracking tools. No ideas were returned for you can't be serious. Google, really? No one searches for anything related to this? Zero. Like nothing. Because when I go to Google itself and start typing, you apparently have lots of suggestions. So wtf? And then when I take those same suggestions and plug them into AdWords, you show me volume numbers for them. Google is AdWords is hiding keywords from you that they know people search for. The why is a little complicated, but basic story. They're not showing non commercial terms because they don't want people bidding on things that will not result in a conversion event for them. And so this data is getting hidden. It's frustrating. I cannot endorse the use of Google's volume numbers as absolutes, but I can barely endorse them as relative numbers. Either these, these are not real quantities. You can do that comparison in Google Trends or you can use other volume ranges. So Moz has a tool for this in Moz Explorer I mentioned Ahrefs before. They have a keyword tool that is also quite good and has some nice volume data. There's also the grepwords API. These numbers, these ranges are about 95% accurate. So 95% of the time a keyword will actually fall into the range that is showing there. So a little bit better if you want the best data though if you want to know absolutely how many people are searching for a particular keyword, you have to bid. There's no choice. You've got to pay Google, you have to bid and you have to look at your impression count. That's the only way you're going to get real data anymore. All right, as we talked about, I wouldn't urge you to Rely on Google Suggestions inside of AdWords search suggest and related searches are smart places. There's also a bunch of other kinds of places you can get it. So people also search for right down at the bottom of the page similar pages that these things rank for Semrush, spyfu Key Compete are great for these semantically connected terms. Topically related searches Questions containing the keywords. Many folks are big fans of the free tool Answer the Public which scrapes Google suggests for question related stuff. And then AdWords is commercial suggestions themselves. There are a bunch of sources for these. I went ahead and listed them, but it makes Moz look good, which is really embarrassing. So we're going to skip through this slide very quickly. Sorry about that. It just happens to be the case like all those other tools are going to have all those other blocks filled out very quickly. Sorry. To a person who was trying to take photo, it'll be like it's in the slide deck. All right, number three, we need to use searcher intent and we need to understand SERP features in order to break into some of these more interesting results. So here's one of my favorite interesting results. I'm embarrassed to pronounce this in the northeast it's a na Easta na Easta Na Easta. There you go. All right. By the way, how cool is it if you worked at Google and they were like ok, we want you to work on the disaster types. Look at that. Disaster types, that's just awesome. I would so work on the disaster type knowledge results. So these different kinds of results appear all over the place. We talked about how, you know, it's only 3% that have the 10 blue links. SERPs like these are what you're much more likely to see. These types of results get much lower click through rate. So we know, you know, we try and estimate this but we might say like a result like that probably has only 24% of all its clicks. You can see here, right? Only 24% would click on that, that or that. Oh, this is a YouTube result. Never mind. Just this one. Low, low click through rates. Right? Really, really frustrating. Very tough to optimize for. Infuriatingly, Google took away the ability for us to be in video results. It used to be the case you could put a Wistia video, you could put a daily motion video, you could put a a video from your own host onto your website or your web page and you could get this rich markup. Now it's YouTube or nothing. YouTube's the only one who gets this Anymore, sometimes, very occasionally. A couple other sites too, but mostly YouTube. And so yeah, Vimeo occasionally because of that. Right. Like we have to think about when do we want to use YouTube, when do we want to use our own sites for things? Same by the way is true for mobile and the app stores. So if you are searching in Google Mobile, there's no way for your website to come up in apps results. You have to make an app and get it into the app store and then Google has to make money or Apple has to make money on your app. That is kind of the way it goes. My recommendation here is you should take a look at all the keywords that you care about for your marketing and, and you should analyze all the features that appear in those results and then you should make a decision. Should I be doing video? Should I be doing image SEO? Should I try and get my site into Google News? Do I need to start playing in the local results? Do I need to start playing in the knowledge graph and one box results? Do I need to think about how I can be in instant answers? If you don't ask these questions, you will lose out to competition. Who does? I think the verticals and SEOs that you need to optimize for seem like a large list. There's only like 18 different types of results and really only about 6 or 7 of those that matter tremendously. Images being a big one. But this can be quite powerful. And image SEO is not an impossibility. It is just as doable, just as investable in as web SEO. It's just that far fewer people are investing there, which is awesome for us because it means more opportunity with less effort. That's a beautiful thing. If there are certain searches to target, there's a couple things that you can do. One is you can aim to try and influence search. Suggest this can be done through branding and offline campaigns. Search behavior type campaigns. You can see a lot of like big brand manufacturers, cars in particular going after this, but radio and television advertising going after this. Ways people talk about a product going after this. And you can also aim to be in those right places when people get there. So Google often with their results leads to another type of search and then you can get into them. This is not a comprehensive list of all the places you might need to be for all these verticals, but you get the general idea. All right. Google's also gotten really good at understanding searcher intent. So if you're looking for, as I was, what's that book that my grandmother recommended to me about the Jews who were living in Alaska. Hey, it's the Yiddish Policeman's Union by Michael Chabon. Oh, there you go. Thanks Google. Now I don't have to embarrassingly call my grandma and not remember. What about that famous circle of big rocks? Nailed it. Place with all the food stalls in Seattle? Pike Place Market. It is not Pike's Place Market. FYI. If you visit Seattle and notice locals getting annoyed when you ask, we love to help people find it. Big underground wheel for science in Switzerland. Large hadron collider. Like they know by the way. I was too embarrassed to put it in, but I did this morning screenshot and almost put in evil orange asshole taking over us and they got it like they nailed it. So back to SEO. Back to SEO. I will not let my anger rule me. Keyword matching not a competitive advantage. Not really anymore. So you want to take big orange guy and you want to put that in your keyword so that you can rank for that. But I'm sorry to report that the only thing that will let you do is give you a chance to rank. It will not. It's not that it won't help you rank, it just can't make you competitive. Like that is table stakes today. On page SEO is not satisfied anymore by just raw keyword matching. One of the things that we've seen is that Google is much more sophisticated in their analyses. And so we try to do things like extract out what are the terms and phrases that pages that rank on page one tend to use that pages further into the results don't tend to use and show those to folks that tends to have a good result. Google's just a ton smarter about the way they interpret content and which content they decide to show. It is still wise to use keywords. It is absolutely wise to use keywords. I am not saying you should stop. In fact, if you stop, if someone tells you keyword based SEO is dead, don't do keyword matching for on page anymore, they're dead wrong. That is not the case. This will help you get higher click through rate. If someone searches for home prices or home values, you should have those in your title. That's not just going to help you have a chance to rank, it'll help you earn that click. But Trulia and Zillow both know that when someone does that type of search, they have specific desires like they want something from those results. They want home prices, they want them by neighborhoods with trend data, they want zip code filters, they want a zoomable map. And look, they deliver on that experience, that searcher experience of satisfying what people want. That is a huge and critical part of SEO today. And unfortunately, it is not in a checklist. Right. It's not a. Okay, yes, I put my keywords in the title. I got it in the meta description. I made sure I have a clean URL. I, you know, it's like, oh, wait, did I make sure that my content matches the searcher's intent to the best extent possible and met their conscious and unconscious needs? Oh, that's quite a checklist item. This is how I think about on page SEO. You do need to continue using intelligent keywords in your page title and your description, your URL. We need to think about related topics and the use of those, not just in the content itself, but in the subject matter and how that's describing it. We've got to be serving keywords that are matching intent together on one page. We need thorough answers to the searcher's query keyword being thorough there. And we want to try and provide unique value above and beyond what other results in the search engine provide today. So if you go to page one and you say, oh, someone's already providing this value, how do I give my searchers that? And more. Whiteboard Friday that just came out this morning is about how we've been seeing a bunch of content comprehensiveness outranking everything else, even sites and pages that have more and better links, which is crazy. That is totally new. But that is a chance. That is a chance for all of us who look and we see, oh, my domain authority's not high and I've been having trouble getting links. How can I possibly compete in SEO? Better experience will actually give you a chance, which is kind of cool. So more stuff in there. All right, number two, what are we doing? All right, we've got to create a link strategy that scales with decreasing friction. It is absolutely the case that links still matter. They are still hugely important. But if you want to do link building successfully, you need a strategic roadmap. I'll show you what I'm talking about. It starts with goals, an approach, the tactics you're actually gonna take, and. And the KPIs and metrics you're going to measure. This is like how you approach all projects. I know, but link building has not classically been this way. So I'm going to show an example. This is Book Surf. They are a Turkish website, a Turkish startup that helps people, specifically in Turkey, get access to books. Books are very expensive and hard to reach, especially for a lot of folks outside of Istanbul. And so these Book Surf guys started. You can see I really like them because look at that amazing mustache. It's clearly these are my people. So Book Surf's link strategy might look something like this. They're going to try and rank specifically for book names in Turkey, right in google.com.tr and they're looking for that in their other operating regions. They're going to encourage their members, people who join the site and who are lending out books to link to their profiles, link to their favorite books. They're gonna try and partner with retailers and authors to earn those. They're gonna nudge people to give the link during account setup and upon completion of the activities. And then they're gonna measure that based on links they get from members, Links and mentions, ranking for key terms. This is kind of how you build your strategy. And then as you start to execute on it, you might start changing your tactical initiatives or you might decide, hey, we gotta throw out this strategy entirely, right? So that is okay. The other thing we need is buy in on experimentation. Link building is a very experimental world. If you don't give it room to try and fail, you will never have great success. The best link builders I know failed for years on tons of projects and then became great. You also need expectations on time frames. So more so than even content marketing, link building and SEO takes a lot of time. And there's often this gap of disappointment between additional work going in without additional ROI coming out. And if you're not prepared for that, you're going to cancel the project before it ever turns into what you want it to turn into. All right, so I think it's okay to balance some long term stuff, right, that have high upfront costs and slow to show roi, but they earn links while you sleep and they have no spam risk with some short term hacky link building stuff. Like you kind of are paying in time or in money to a link builder as you go along. You don't want to pay directly for links, but they can show some fast roi and they can, you know, effort in means you can get links out, but they sometimes can have spam rigs. If you play it right, you will be fine. Right? But this is sort of classic. A lot of classic link building falls in this bucket and a lot of what we call like the content marketing based link earning falls into that other bucket. The best link builders I see, they focus on the flywheel. I'll show you what I mean. So Moz has a flywheel and I'LL use us as an example. We do keyword research, we gather industry information, right? I have a bunch of discussions with all of you in the hallway after this and you give me a bunch of ideas for posts and things that I should write and talk about that will hopefully be helpful to other people. We publish that content, we push it to email and to our RSS subscribers. We promote it via our social channels, right? Twitter and Facebook and LinkedIn and, and even Google still. We earn links and amplification from that and we grow our social channels and our networks. We get a few more subscribers via rss, a few more subscribers to our social channels. We grow our domain authority from earning a few more links to our site and to our content, we earn that search traffic and referral traffic over long term. So there's like that initial spike, but then there's the long tail of traffic that continues to a piece of content and then we can rank for more and more competitive keywords over time, right? And this flywheel is a beautiful thing. Now, in 2006, if I published a post, I was lucky if that post could get any amplification whatsoever. In fact, it almost never happened. I remember cheering, cheering, walking up and down like, you know, walking back to my apartment. Not my apartment, Geraldine's apartment. I couldn't afford apartment at the time. I was living with my girlfriend. Now she's my wife, but so she. The, you know, the thrill of earning one link to one blog post was incredible. And then 2016, I mean, it's embarrassing to say, right? But I hit publish and links sort of just come in and lots of them usually, and that's a wonderful thing. But it's also 10 years, 10 years to build up that flywheel to work in this fashion. We are a business, clearly we are a business built on content. Dollar Shave Club, good example. Their flywheel is built on press and media, mostly around their creative advertising dribble built on a user generated content community. Almost every single flywheel you will see finds points of friction, right? Someplace where they are stuck, where they can't make the flywheel work, where it's just hard. That is a great place to start implementing those hacks to say, hey, we need to give our flywheel a little more energy. Let's find that point of friction, let's inject those hacks. That is not necessarily evil. It doesn't necessarily have to be spammy and it doesn't have to be without value. As long as we have a strategy, we can apply our links there, right? That is the model that I think modern link building works in, and it uses both the new world of link earning and the old world of sort of classic link building. Last one here. All right, we got only a few minutes. We're seeing this HubSpot. I think you all recently, you all who are in the room recently did an experiment where you did some searching and saw that by searching and clicking on things, you could change their rankings, which is pretty cool. We've done that at Moz. I did that last year at Inbound. That was pretty fun to see for those of you who were around, maybe that was two years ago. So rankbrain is one, one part of this. Rankbrain sort of interprets a query and says, which algorithmic elements should we apply to this particular search? And RankBrain is also looking at queries and saying, oh, a lot of these all have the same intent. So let's mostly show the same results because we're mostly using the same signals to rank these things on. And one of the ways that Google gets at this is through user and usage and engagement data. I hate bullet points. I apologize for putting any bullet points in a slide, but I'm copying Paul Har's slide here. These are his bullet points, not mine. He's essentially showing that Google says a result is good if lots of people click on it and engage with it, and it's bad if they don't. So here's Jeff Dean, one of the geniuses in engineering at Google think he's a Google fellow. And so he might say this is a really good search result, right? People rarely go to page two, they rarely click the back button. They're generally happy with the results here. And this one, this is a bad search result. Like people who get here, they are unhappy, they click back, they change their search query, they click on page two and page three and page four. Anytime they click a result, they come back. They're just not satisfied, right? And so they can take that model and then basically see when people are unhappy. You might have seen this. Have you seen this in mobile? You do a search, you click the back button and Google starts to show you, oh, maybe you should search for these other things instead. They are measuring it and they are using that data. So I think in the future we might have Google's moving this deep learning, machine learning model. I think in the future we might have the fact that all the ranking factors and all this training data and this learning process is, is not built by hand by software engineers. It is built by machines. And Machines that don't even necessarily need to get the inputs from human beings, they can figure out what should rank and how it should rank based on the data that they get from all of us. Google's leveraging this deep learning stuff. Despite this literal quote, we don't fully understand our own algorithm that we built. But that's OK because it works. FYI, this is how the Terminator comes about in case you were wondering. So I think we might see in the future and even are seeing today this world of non universal algorithm inputs. You remember this chart that Moz kind of does every couple of years where we collect all these ranking factors, then survey people. This I don't know if this world will still work. I don't know if a world where I say yeah, you know. Tends to be the case that domain authority matters this much. Tends to be the case that page level matters this much. I don't know if I could tell you that in the future. I think in the future it might be. Well for some queries this doesn't matter at all. And for some queries this matters a ton. And for some queries this matters a lot. That is a world that we are starting to live in and I think will live in more so in the future. If you are interested, you can read Steven Levy who wrote in the Plex, did this article on Backchannel all about Google's deep learning stuff. How do we do this? We as marketers are going to have to focus on engagement. We are going to have to focus on this engagement reputation that websites build up, right? Because the quantity of pages that are earning visits divided by the relative time and engagement that people have with them is going to give Google a sense for whether your site satisfies people. And if you don't, they will put you lower in the results. Your competitors will rise. When you go into your analytics and you see pages that are not doing well receiving search visits, you gotta call it right. You gotta say hey, these pages are sucking in the search results. People are unhappy, they're bouncing, they're leaving. And that is gonna hurt these classic click through rate. We gotta find a way to beat our averages in order to win. And to do that I think we have to serve multiple searcher interests, not just our own. If you are thinking how do I sell these people sous vide cooking machines? You're asking the wrong question. The right question is how do I give these people everything they could possibly want when they perform this query? We need to empathize with searchers and give them the content that serves those needs. FYI, nobody, nobody who ranks on page one for sous vide cooking is selling anything. No surprise. And that is not because they don't have the links to do it. They do. There's plenty on page three and four to do this. I have a few recommendations. Obviously authoritative content that is fast, that delivers a great experience on every device and compels visitors to engage share in return. And certainly no offense to Neil Patel, who's a very successful marketer, but maybe you don't need four pop ups on every page. Maybe just one seems good. No, I don't want to become a better everything. With that said, one of my big takeaways is not just great content, but but very focused, very empathetic content. And I invite you to learn more about that. It's time of big change. I think that in Google and in SEO we have the tools and the skills to compete. And that is a beautiful thing. And by paying attention we are so much further ahead of the field than the millions of folks who are trying to compete with us. I. I wish you all the best of luck. I look forward to chatting more and you can find the presentation online. Thank you very much.