STATE OF SEO, GOOGLE, AND TIPS FROM HEAD OF SEO AT YOAST

Robert Hansen

But also in the marketing world, you're like, “Yeah, that means this content’s never getting refreshed. That means Google doesn't care as much.” So it's very interrelated, which I find really fascinating.

I always feel like the odd duck out at this conference because I'm the only security person here. But everyone has a tangential interest in security because they want to make sure their sites stay up.

Jono Alderson

Yeah. SEO in the last, however many years, decades, has very much progressed from tactical hacking, for want of a better phrase, how do we beat the machines? How do we find which levers to pull? From starting there towards very much now brand building.

Yes, there are still tactical action opportunities. But really, the long game and the big wins are I'm going to grow a brand and create an empire and build a castle and make my website bigger and better and faster and stronger over time.

A huge amount of that is, you're building a castle on top of architecture. And if somebody comes in and you've got a big security issue or an accessibility issue, it rips out those bottom bricks. Your site crumbles, your reputation crumbles, your user preference crumbles, and your SEO is all intertwined with that. Yeah, these things are very much tied.

Robert Hansen

I don't think people really understand how interconnected things like Google's ability to tell that somebody bounced back to Google and how that's going to affect your rankings. Those types of user metrics, Google is paying very close attention to those things.

Jono Alderson

Yeah. There's huge amount of speculation on how that might work. We're not entirely sure. But we do know that broadly, the interactions that people have as they're going through search on their journey, nobody searches for a keyword. Consciously evaluates the results, and then clicks one.

We all bounce back and forward. We visit five different websites on three different devices over how many different days, and all of that's monitored. And yet, you can understand a huge amount about, “We have ranked this site first, yet everybody who clicks on it immediately comes back to the search engine and clicks the one underneath. Okay, that's wrong. Swap them.”

We’re losing tangential evidence of all this sort of thing happening all the time at scale. The user experience and security and people's trust in sites in particular are really key.

Robert Hansen

Site speed is another really good indicator of that. From Google's perspective, this makes a lot of sense. What they don't want is for users to go to a site, think it's broken, and therefore think Google's broken for having recommended it.

Jono Alderson

Yeah, exactly. That's a really difficult relationship that, yeah, Google's curating these websites. Your experience on that website reflects on them. And if that goes wrong, then people imagine going and using Bing or some other search engine that no one's ever heard of.

Robert Hansen

Right. I've literally met people who truly believe Google is the internet, “This is my internet.” If anything goes wrong, they truly believe that Google is at fault. They don't give the brand any sort of like, “Well, eBay is messed up today.” They're like, “No, it's Google's fault.” So Google really has to penalize you just to make sure that their customers are served well.

Jono Alderson

Yeah, they can't afford to have a bad inventory. And I think this is something that the SEO industry struggles with. There's an increasing trend where Google is extracting the content from websites and serving it in the search results.

Part of that's for this reason that that is a better user experience than sending a user to a website that they don't necessarily trust or aren't necessarily familiar with.

Even if they are, they still have to learn. There's a cognitive load in, where's the navigation? How do I find this? Actually, Google can do a good job of surfacing it and directing the search results. That is a far better user experience.

Robert Hansen

That is a very good point. Once upon a time, one of our marketing guys left a company I used to work for. He left me 150 different social profiles that I had to go change the password on. These are all totally different kinds of websites.

What I realized extremely quickly is that every one of them fell into super obvious, this is how to do it, or massive clusterfuck that took a lot of work to figure out. Sometimes you had to do it on different websites. Sometimes you had to authenticate with some other token.

It was bizarre, weird, and gauntly. Sometimes it's under me, sometimes it's under profile. Sometimes it's under change passwords. Sometimes it's under settings.

To your point, the cognitive load that a customer has to go through just to even figure out what's on this page and scroll down to the point where they need to be, if a customer doesn't see what they want to see pretty quickly, that's a good opportunity to bounce.

Jono Alderson

Yeah. What we're seeing is there's a mind shift, which people are really struggling with which is, “We've built these empires and marketing systems based on the premise that we can get traffic from Google and then try and convert it.”

The fundamental misunderstanding is that has never been your traffic. It has been Google's traffic, and it still continues to be Google's users. And it's in Google's best interest to not have those users experience friction and to have to click a website they don't understand.

Of course, they're going to extract that content and share it themselves. And we're now facing an existential crisis where there is no good answer to, how should this work? What are the economic models that benefit content producers if their content is just extracted and served into Google?

How do I run a business which is fundamentally based on clicks if nobody's clicking websites and visiting them? There are no answers to any of these questions. But everybody's KPIs and everybody's bonuses and everybody's marketing machines are all based on this. There's a real evolution happening, which is really fascinating.

Robert Hansen

Yeah. We can talk more about that, too. Let's talk about Yoast for a second because I think this is worth explaining. Tell the audience what I think you and I already know about how incredibly pervasive WordPress is and how important Yoast is for the community.

A lot of users are not going to know the word Yoast, but they're impacted by it whether they understand that or not.

Jono Alderson

Yeah, most of the websites you visit as a real human being probably run Yoast. WordPress now powers 43, I think, percent of the web depending on how you measure it. That's based on the top million websites in Alexa’s dataset, which are being discontinued. But the stats looks similar to that depending on wherever you slice.

That's just the top million. So that's also excluding huge amounts of the real web of many small sites in the long tail. WordPress is huge. But it’s open-source. It's very poorly managed. It's very poorly run. So it does a poor job of marketing itself.

If I were to try and describe it, it has become fundamentally the operating system of the web. If you are building anything that looks or feels like a website, it's probably one of the best tools to choose, probably one of the easiest to access news.

It has the largest, readily available marketplace for themes and plugins and filters and stuff and developers and designers and people who can use it. There are other ways to build websites. There are other platforms and technologies. But really WordPress has become homogenous.

I spend a huge amount of my time looking at websites and being on and discovering websites. Nine out of every 10 that I visit, if you view the source code, it says Yoast in there somewhere. One of the challenges of WordPress is there's a fundamental principle that it does as little as possible.

The core heart of WordPress is deliberately quite shallow. It doesn't ship with anything like a contact form. Or there's no inherent security add-on. There's no email infrastructure, there's no marketing tools, there's no SEO.

Core WordPress does some very basic, rudimentary SEO stuff. It gives you the absolute bare minimum.

Robert Hansen

It does your title tag.

Jono Alderson

Yeah, and barely even that. Only you can't differentiate from the name of your post, which is quite fundamental. I don't necessarily want the same words to appear in both those places. That's by design, it's intentionally quite thin.

There is a huge ecosystem of plugins around that if you’re like, “I want a contact form, I want a security plugin, I want an SEO solution.” you shop around and pick one out of the masses.

For a decade, Yoast has been the biggest SEO plugin by an order of magnitude, by several orders of magnitude, in terms of number of downloads, number of installs, quantity and level of reviews.

We are synonymous with WordPress SEM because a lot of people who actively want to market their site at some point in their journey will go, “I have a site, I have a few pages, I want to do some SEO.” They search for SEO, and they find us. Really great.

We have a freemium product and a model and then an upsell to a whole bunch of premium add-ons which unlock more tools and capabilities.

Robert Hansen

Which do what? Give us some examples just so they understand what you're talking about.

Jono Alderson

The free tool does everything that you need in order to compete in Google. This is contentious, but I think it breaks down into roughly three categories. Your site has to be discoverable by Google, which is essentially technical standards.

Do you have a functioning XML sitemap? Are you outputting the white meta tags? Are you aligning with best practice standards? And again, WordPress doesn't do much of that out of the box. Maybe 5% of what you need, we fill in the rest of it.

There’s a huge amount of heavy lifting there to add that capability on. Then you've got to be indexable. The content you're writing has to align with various standards and guidelines. Some of that is technical, and some of it is editorial.

Google will actively not index and consume and rank pages which are not very good or which contain very different text or which don't have certain combinations of meta tags. Some of that we do automatically, some of which we guide you through the process too.

Then you've got to essentially beat your competitors. You've got to have content that's good enough to get cut through. The rest of what Yoast SEO does is essentially-

Robert Hansen

Good in the context of compared to competitors?

Jono Alderson

Yes. Better.

Robert Hansen

But from the users’ perspective, ideally.

Jono Alderson

Yes. A human being has searched for something that your product or service or business either solves directly or has an opinion on or has content that helps somebody on a journey. Is that better than the next guy? Does it fill a gap? Yeah.

That's the rest of what we try and solve. We have a whole bunch of content tools, which do things like analyze your content as you write it, and they'll say things like, “We noticed that you keep starting paragraphs with the same word, which makes it weird to read. We noticed that this section is very long, you should break it up into subheadings.”

Some of these are explicitly SEO stuff. You should probably mention your keyword more. Because if you're trying to write a page which is a comprehensive guide about cats and you never say cats, it's not going to do very well.

That's not about getting the perfect ratio of keywords or stuffing it everywhere. It's about being on topic. Google enforces a model where one page is about one thing. And that requires a subtly different type of writing and content shaping than you might do if you were writing a book or a white paper.

Robert Hansen

So I’ll have to rename my feline forum?

Jono Alderson

But then, you also want to be mentioning kittens and puppies maybe a little bit depending on where you want to go. We give you a feedback mechanism for all of that and say, “Do a bit more of this. Do a little bit less of that. Have you thought about this?” And guide you through that.

Some of that is explicitly SEO stuff, but a whole bunch is readability. Is this text easy to consume? There's some contentious feedback on that because it does tend to make people simplify their content. We think that's good. We say use shorter words, use simpler language, write shorter sentences. It's quite hard.

Robert Hansen

Especially for guys like us.

Jono Alderson

Yeah. I score really badly when I write stuff, where I have to rewrite and use fewer syllables.

Robert Hansen

This had more and more parentheses.

Jono Alderson

So it guides you through that in the hopes that if your objective is to get lots of traffic from lots of people from lots of different places, the more accessible you make that, the wider you're casting a net.

Yes, you might have quite an expert topic and might be quite in depth. But your objective is still going to evolve to a degree around accessing the widest possible audience. So it's quite cool.

That's part of it. We also started to expand out into more general content stuff. We just launched an inclusive language feature, which is really cool, which detects when you're using phrasing that might be offensive.

I won't give any examples because it's going to be offensive then fall into a trap. But then we provide recommended alternatives. We're starting to go broader into not just, are you ticking the SEO boxes? But are you genuinely producing content that's good and helpful and useful and inclusive? And all this other stuff.

We're the backbone of SEO in WordPress. And yeah, more recently, Shopify. In January, we launched a Shopify app that does much the same thing, slightly different treatment and context.

If you're on Shopify, we know your e-commerce. It's a slightly different set of needs and slightly different set of rules, but the core principles are the same.

Robert Hansen

You are the first pure play SEO that I've brought on board to speak. Briefly, I wanted to talk about just SEO versus SEM and just the data and how you see those things.

I did have two other people on; Josh Castell, who did marketing. And another guy who I think by the time this comes to air will already be out there, Elijah. But I wanted to get your take.

First of all, before we jump into that, what does SEO at Yoast look like? You're not trying to make Yoast rank.

Jono Alderson

A little bit. Yeah, that's not my main focus. But to a degree, we have to dogfood. Nobody is going to buy SEO software from a company that fails fundamentals. Because we're so big and because we're quite opinionated, we paint a bit of a target on our backs.

If and inevitably when we get some stuff wrong, we take a lot of flak for it. I do spend a little bit of time making sure that we're following our own rules and guidelines and getting stuff right.

Robert Hansen

That’s cool. I didn’t know that.

Jono Alderson

Not officially, but it's quite interesting to sniff around that.

Robert Hansen

Yeah, I bet.

Jono Alderson

One of my favorite things about SEO is nobody really truly comprehensively understands how it all works. Not even Google. The black box and degree of machine learning, algorithms within algorithms feeding other algorithms, nobody really gets it.

I don't really believe in experimentation in SEO, but there's certainly a degree of constantly iterating and improving and refining and checking. Things change, people tweak things, stuff gets missed all the time in the same way it does with any company. So I chase the website around and have a look. But that's not the main job.

Robert Hansen

Okay. What would your main job be then? Defining new strategies for the-

Jono Alderson

Products, product features, and strategy, which are the many things we could build or we should build, and technically how they should work and what they should look like.

In a world without Yoast, most websites would need to go through a process where they said we need to go and understand all of Google's technical requirements for things like XML sitemaps, canonical URLs and then develop an implementation for our website, which uses our business logic. That takes a couple of weeks. Fairly straightforward.

The challenge we have is running on 14 million websites. We have every possible combination of not only our own software but a billion different hosting environments; PHP versions, Apaches, engine axes, SQL, different database formats, sharding, internationalization. Nightmares.

Then alongside that, a WordPress has something like 50,000 plugins in the repository. And there is no dependency management system. So I can write a plugin right now that breaks your plugin. And if I run them both together, my site will fail others.

We have a huge amount of time making sure that the rules and the code that we're writing and the features we're developing interact nicely with an infinite number of possible combinations.

Then there's a theme ecosystem, which is alongside that. Depending on the code that you're using to display your pages and your content, there's further dimensions of potential conflict.

As an individual business, saying we should implement canonical tags is pretty straightforward. I've got a 6000-word document that has to consider every conceivable eventuality in an edge case, and then we've got to write code to implement that.

A big part of my job is going over the next 12 months, “What are the features we're going to build? How should they work? What's the test spec? And how do we make this perfect? How do we solve this problem?

I think it’s a really enviable position for an SEO to be in, to be not just patching and tweaking for one site but to be going, “How do we fundamentally solve, conceptually, these types of requirements?”

In some cases, that's really tricky because the specs themselves aren't great. And they don't scale infinitely.

Things like if I've got an international setup and I need to describe the relationships between all of my different versions of my pages in different languages, that doesn't work when you hit a certain scale. And it doesn't work when you have other dependencies and interactions.

We got to find workarounds and clever solutions. And some of that is guiding the user to make certain decisions. But we don't have any context. Some of it is choosing what not to build and where to draw the line. It’s a fascinating challenge.

Robert Hansen

14 million websites, if you mess up, you're going to get some phone calls.

Jono Alderson

Yeah, which does happen occasionally. Not only 14 million websites, but as the mechanisms by which we ship our software, we publish an updated version that lands in the WordPress ecosystem.

People either have their sites automatically update, which immediately breaks them. Or they go and hit update, and they download the latest version.

If it's broken, there is no mechanism for them to revert unless I've set one up and I've got a backup system, which many of those websites for small businesses won't. And the only fix is we've then got to ship another version.

As soon as it's out the door, if there's issues with it, it's a huge problem. And even if we try and revert, they've then got to get the latest package, which if their site's broken isn't going to work very well. It's an absolute nightmare.

We spend huge amounts of time and resources testing and triple checking, smoke testing in different environments, where our code for our test scripts is typically about three times longer than the actual solutions they’re having for really simple stuff. Yeah, it's pretty huge.

Robert Hansen

It's funny you say that. I come from a waterfall product management background. A lot of people were like, “Waterfall, that's so old school. I can't believe you still do that, blah, blah, blah. Is it eBay?” But eBay was a similar problem. If you shipped bad code, you have now directly impacted 150 million items.

Jono Alderson

Every minute that goes by where that's still broken is-

Robert Hansen

And it may stay broken for weeks until you get this thing rolled out again and fixed and whatever. It was closer to sending a man to the moon than it would be a typical software. It’s like physically shipping software.

Jono Alderson

Yeah, we’re sending our CV out, essentially.

Robert Hansen

Yeah, almost. If you screw it up, you're not going to be able to help them. They are on their own.

Jono Alderson

That’s right. For the first time, we have competitors now in a way that we've never really had. I won't name them because they're awful, and they steal our code and rebrand it and present it as their own and say nasty things about us.

For the first time, there are alternatives. If we make a big mistake, we'll slice users straight out. And our market share will drop. The care we have to take to make sure that this stuff is infinitely resilient is fascinating.

Robert Hansen

I bet. Okay. How would you define SEO and SEM and the difference? I asked Elijah May the same question. I'm curious. I want to make sure that the audience is really clued in about how an actual professional SEO sees those things.

If you were a business, which one would you do? Or would you do both? Or why? Or how or whatever?

Jono Alderson

Wow, tricky. Every SEO you ask will give you a different answer.

Robert Hansen

Which is why I think it’s an interesting question.

Jono Alderson

SEM, I think that's the easy bit which is you bid for traffic. You access a marketplace, unlike on eBay, and you blind bid for traffic with some degree of intent to a landing page that you've crafted.

Depending on the level of competition and the kinds of phrases and intensity you want to target, you're going to pay more or less. It's going to work better or worse. And you can scale that up on demand or you can turn it off or you can pivot strategy immediately.

Very nice on demand. Good if you're doing short term strategies, if you're not interested in long term brand building, if you have something that's time-sensitive and you need to crank up the traffic and you know that you've got sufficient throughput that you got to make enough revenue, etc. Great, really good.

Almost certainly, it makes sense to run in parallel with SEM. Let me take a step back. Actually, most brands and by extension, most websites will, to a degree, describe the thing that they sell. I am a dentist, I offer dentistry, look at my dental products. It's not interesting, it's not compelling. It's not differentiated.

You can run all the SEM you want, and people who want your dental services might find your page, and I'll come through and convert. The SEO challenge is I need to be able to reach people two months before that point when they go, “I have had toothache for two days.”

They are starting to explore the problem space. They're not looking for a solution, they're starting to explore the problem space. To win the SEO game there, you need to have content and reputation sufficient to convince an omniscient AI that your webpage is the best possible solution for the problem they're having there and then.

It isn't necessarily going to be, “I want to buy dentistry.” We were talking earlier about this is Google's audience and Google's users probably. The best outcome for that user is a guide and a video and a resource that helps them self-diagnose that might say, “You know what, you don't need to go to the dentist. Maybe you should just floss and then check it in a week or so.”

SEO is doing that at scale solving Google's users problems before they get possibly also at the point where they're going to spend some money. Great, much more competitive. But the real opportunity is much how the funnel is doing that scale. Then also finding a way to connect that to your commercial engine.

What you might end up doing is at the bottom of that guide say, “Here's our newsletter, here's our white paper, here's our webinar, learn about teeth.” That's a strange topic. And it's quite an alien way of thinking. All marketing is grab attention trying to sell. Actually, what we're now saying is intercept problems and try and help.

Robert Hansen

And continue to help.

Jono Alderson

Yes, and help people who probably will never buy from you. But then you reach a million people. And if 0.1% of those go, “Oh, okay. Now I do need to go to the dentist. I remember that that one was really useful.” You’ve built a positive brand association, and you hope the conventional marketing and advertising rules apply. And you capture some value out of that.

It’s hard. You can try and compete with the same commercial. But those kinds of terms are much more competitive because everybody wants the easily converting stuff, it's easier to measure, it's easier to build business cases. There is already a lot of content out there, there are already a lot of websites.

You need to be producing resources which are 10 times better than the best people in the space already, which is a huge ask, and then to also scale that. And again, this is quite an alien way of thinking for businesses. None of this sits in conventional budget spaces. None of this is well-understood.

Robert Hansen

It's not like a physical mailer I would use to send. A lot of businesses totally get that like, “Well, I'll reach a million people. I'm sending out a million envelopes.” This is like change some stuff on a website, and suddenly 10,000 people will show up. It's like, “Wait, what caused that to happen?”

Jono Alderson

Yeah. No matter how much you try and test and evaluate and reverse engineer, you will never get to anything close to understanding, “I made this read, and therefore I got more.” No, you have no idea. Because also, you don't know what the competitors do.

You don't know how the landscapes changed. You don't know what the impact of the weather was. You don't know how the fundamental size and shape of the internet is changing. Maybe Wikipedia did something. You have no idea, and you can't isolate any of those enough.

Robert Hansen

We had one presentation today. Literally, the weather changed the algorithm, the output. And then we spent weeks trying to figure it out. Turns out it was the nice weather in Germany. That's where we lost a whole bunch of traffic.

Jono Alderson

There’s a huge subset of the SEO industry, which is obsessed about this testing. They'll go, “I have discovered that if I make this thing blue, I will get more traffic. And we've tested it.” You haven't. None of them have. I think there is no way you ever get to that kind of clarity.

We have broad direction. There's no surprises in that really. Again, it's about, is your content helping Google's users to solve Google's users problems? Incidentally, yes, you can try and upsell off the back of that if you're graceful. There are no magic tactics really, which is a shame.

Robert Hansen

I want to talk a little bit about the history of how we got here because I think this is a pretty interesting long tail of a conversation. From my background and where I got involved was right after GnuHoo turned into NewHoo and turned into DMOZ.

That's right when I was getting involved in understanding anything about search at all, which is a long time ago. For those who don't know, that was right around the time they were bought by Netscape.

DMOZ’s idea was, “Why don't we just get a whole bunch of people, moderators who just are really passionate about whatever it is, dogs dentistry, whatever. They'll curate as many things in that category as they can.”

People who have stuff in that space will go to DMOZ and say, “Hey, I have a website that's about dogs or dentistry. Please put me in your index.” Then they go there and they go, “Oh, this is for dentistry in Munich.” It's like, “Okay, go to the dentistry.” Maybe it's location-based, now Munich.

Now you have a link there. And it's alphabetized or whatever. That was hugely important for companies like Yahoo. And eventually Google, they just basically scraped that like crazy.

Jono Alderson

Definitely, Google and, in fact, Yahoo had a patent around trust rank quite early on. The premise was the number of hops you were and in terms of links from an authoritative resource was quite a key indicator as to your trustworthiness.

If I'm linked directly from DMOZ and probably a legitimate business, if I have to go through 100 steps, I'm probably a shady back alley down the history of a legit one. DMOZ was huge. Then Google, not quite so explicitly used it. But definitely, it was part of that consideration.

Robert Hansen

There was this time period when there was this concept called LUMP, which was an acronym that stood for the link. I think it meant how the link structure works or links to it. I guess it's probably it.

U would be the URL, the actual structure, did it say the exact match of dogs or whatever? M would be meta tags back when meta tags really were a thing, meta keywords and meta descriptions.

Jono Alderson

Meta tags are more of a thing than ever.

Robert Hansen

Oh, they’re back?

Jono Alderson

Not meta keywords. But yeah, there’s a resurgence.

Robert Hansen

Okay. We'll talk about it, too. Then P would be placement on the page, where your content physically falls.

Jono Alderson

Yeah, because something in the middle is more important than something in the footer. That's reasonably easy to determine.

Robert Hansen

Obviously, if something is the H1 tag at the top of the page, that's probably more important than something that's in the middle page. How did we get from LUMP to where we are today?

Jono Alderson

Because of the kinds of challenges that were mentioned around, let’s say you're not really fitting anywhere into an organization because nobody's interested in producing helpful content. They want to be selling stuff. It always existed at the fringes.

It was always exclusively tactical. It was, what tactics can we apply in order to try and rank higher to get more visitors? It never tried to address the fundamental challenges of, how do we make my website better and more helpful?

That manifested in SEOs essentially breaking all of the things. And one of the first things that they broke was directories like DMOZ. There was a realization that if there is some intrinsic value in being listed in these types of directories, great.

Robert Hansen

Let's put ourselves in there.

Jono Alderson

Yeah, and then let's have agencies that their clients pay them to put them in for them. Let's have services and middleware that do the same thing. Let's scale this until it breaks. And lo and behold, it breaks.

Robert Hansen

How did we get from LUMP to, let's say, PageRank or RankBrain?

Jono Alderson

As we were saying just now, one of the challenges that SEO has always had is that it's a quite hard conceptual model, the idea that it doesn't work like other marketing channels. You can't just say, “Let’s put adverts or websites or stuff in front of users and hope that some subset of them convert.”

It's never sat anywhere comfortable within organizations. It's always been extremely tactically-focused, certainly back in the olden days, maybe lesser now. The whole model then was we have this weird external team sit over here which will give you a bit of budget.

Hopefully, you can drive some traffic. And wow, the results are amazing. But all of that is driven from essentially tactical hacking of whatever today's opportunity is.

The first big thing that really worked like that was directories like DMOZ. To the point where as soon as it becomes apparent that there is inherent value in being listed in these directories because Google is looking at them, Yahoo's looking at them, they have some impact on where they decide which sites rank.

Of course, everybody goes, “Right." Everybody puts all of the links in all the directories, and services evolve, and agencies evolve and the links start getting traded and the inherent value increases.

And then, at some point, Google and Yahoo and everybody else goes, "Ah." Directories are now fundamentally worthless. They have been utterly polluted by SEO tactics to the point where we are going to completely ignore them.

We live in a world where actually the idea of directories and D-Moss would be really interesting. I would really like a human curated service that helps me categorize and explore stuff. They just don't exist anymore because SEO killed it.

And then, went on to do the same thing with blog comments. So, exactly the same thinking. It turns out that if one website links to another, that's a sign of endorsement. Google will spot that. And if it's a bigger website or a popular one, that's a bigger sign of endorsement. It's going to affect my site ranks.

These sites have blog posts and comment forms. I can go fill out a comment. I can say that I'm from jonasdentistry.com. Great. That's a vote. And again, exactly the same thing unfolds. We do that at scale. All the tactics get rolled out, agencies evolve all the spam, and now nobody runs comments anymore.

And this has been the pattern for SEO for 15 years or so. It hops onto the latest tactical thing and it milks it to the point that it dies and then you hope that you survive enough to do the new thing and all the agencies rebrand themselves.

And today it's digital PR. And 2013 it was infographics, before that it was guest posting, thing after thing. We're finally starting to see the industry evolve and mature now with things like Rank Brain.

Essentially Google's increasing intelligence, not the right word, but Google's increasing ability to evaluate the actual quality of the underlying thing as opposed to just scrape the words and count the inbound links actually doing a much better and better job of saying, is this a good resource?

And the more, the closer they get to that, the harder the tactics become. I'm getting to the point where I don't think there are many SEO tactics. I think there is only strategy and that essentially boils down to have a better website that is more helpful and more useful for more users.

Robert Hansen

So, we used to have a lot of like Viagra spam, like all over the SERPs, Search Engine Results Page is what that stands for. And now you really don't see that. That has largely gone away.

Jono Alderson

Yeah. They've done a phenomenal job of clearing up the bottom 30% or whatever it is of the internet. They've got very, very good at excluding things which are definitely bad. And there's two interesting components to that.

I think one is we are hearing increasingly from them that they are conscious of storage space. So, the internet is now very big. There are very many pages and they are actively choosing much more often to not index things at all. It is not in their interest to create a database of everything on the web when so much of it is Viagra.

So, they're getting much more nuanced in what they choose to exclude. And the barrier for that is still quite high. Most normal websites, normal pages won't struggle to get into Google. If they are, there's something else going on.

But yeah, they've certainly sliced off all the bottom. In parallel to that, there's some really interesting discussions happening around what's left. And the top 10, 20, 30 results in any competitive industry tend towards being over-optimized. And as you start to click through these in Sydney, you see the same patterns.

You can see that they've hired an SEO agency, they've been given a brief to write 500 words of copy, and they're going to structure it in this way and they're going to use the keywords just so. Maybe they've used Jost for better or worse. And all of these pages are not quite helpful.

They are designed and built to be just good enough to outrank their competitors and to be good enough to be in that spot. But they're always honey pots or lead captures and they're never really truly helpful.

And there are many sectors and industries where the first two or three pages of Google are just going to be content that's been designed to rank an SEO.

And I think where Google are going the moment is trying to solve that problem. Which was about two weeks ago we had the helpful content update. It's very rare of them to name updates because they launched thousands a year. But this one was a milestone.

And the premise is that those kinds of pages shouldn't rank. And there's a kind of rebellion, there's a whole bunch of users saying, "No, we should abandon Google and we should use Reddit for everything because that's better in so many ways."

But if you're looking for authentic, truly helpful content, like the dentistry self-diagnosis thing is probably going to be a better result than the first 20 pages of dentist sites which have generic blurb, which is all just recycled.

So, I think we're almost at the final step of this evolution where SEO truly becomes about have I produced helpful content that deserves to rank and ups the user rather than just have I completed whatever set of tactical requirements I need to outright my competitors this week before we break whatever the latest trend is.

Robert Hansen

Yeah. So, I would like to think of it as, so Google has a product that product is ads. And SEO is the tool they use to sell those ads. If SEO didn't exist and they didn't give a certain amount of latitude SEOs to do what they need to do to make things rank better, then there would be no ads.

The ads would not make sense because you'd say, "Oh, there's nothing I can't control anything about anything that's going on here."

Jono Alderson

And there's a huge amount of content that doesn't align well to ads where it's informative or educational and you want to get the users in, get them hooked to using and relying on that. And then, you can do the commercial ads alongside that.

And then once they've built those behavioral patterns out, it just also happens that you're searching for things which have direct commercial outputs and can be easily advertised and converted.

Robert Hansen

Yeah. I think people largely think of Google as a search engine, but it's really an ad engine.

Jono Alderson

It's ad engine, definitely. That's where the money lives.

Robert Hansen

You and I long ago at another conference had a conversation that always stuck with me. I think we both were circling around the same idea for a while and then we just both kind of figured it out at the same time in the same place that maybe there is no way to do SEO anymore.

For those who don't understand, there are so many sub-engines within Google that do tiny little things.

Jono Alderson

Algorithms within algorithms.

Robert Hansen

So, you were describing that, but like specifically let's say something is an looking for what your Better Business Bureau report is or whatever, and it suddenly starts using that as a key signal or something. We're just looking at your website, you haven't made any changes at all and suddenly you're going up or down or whatever.

You're like, "What is going on?" But they only do that in three categories and you're just happen to be in the one.

So, all these other people are doing these tests can't see it. And they're like, "No, it's not that because I went up and down in Better Business Bureau and it didn't matter." And so you're sitting there trying to optimize one thing and it's actually something over way over here that's happening.

And so, SEO in the traditional sense that it used to work where it's lump where I have control over all the variables, it literally may not exist at all anymore. And if it still exists, that's great, but it might only be a temporary situation.

Jono Alderson

Yeah. And I think the model I was describing earlier kind of fits with that. The kind of base level of have I met best practice enough to be eligible to be discovered? Anything beyond that? Yeah. It's all going to be black box and it's going to be a huge, really hard to understand and manage that.

I think there's a bigger challenge here, which is all businesses forever of probably all types have operated on some flavor of MVP and go, "What is the minimum viable effort, cost, energy, time that we have to invest in order to get to market, to launch our product to whatever?"

No business ever gets beyond that because if you aim for 80%, you inevitably only ever hit 60 and most things are broken and a bit rubbish. SEO requires you to aim for 100% and hit 90. And the one brand that does that will win and everyone else will lose.

And just businesses are fundamentally not set up to care about quality. Like how many businesses have somebody internally owns quality? How many businesses even have a concept of quality? And yet that's the requirement. It's never stated quite that explicitly though.

I say that Google has a document that they give to the human quality rating, testing guidelines. They employ 30,000 humans who essentially vet and validate whether their machine learning systems are producing the right kind of results by looking at results.

The word quality appears in their 300 times. It's not a word I ever hear SEOs using. Yet if we start to say that the way to win is to produce better, more useful hot content at scale, et cetera, the word quality seems to me a lot more sensible and universal than the word SEO. And that's increasingly what it's becoming.

You can't have a bad website rank very well, unless everyone else in that sector is bad or you are exceptionally good in some way. And that isn't about finding the one magic tactic. It's not about the better Bureau results. It's not about updating your copyright.

It's about making a thousand small things better and slightly better than your competitors.

Robert Hansen

And you might stumble upon the right answer by having a good better Business Bureau report because you have a good website.

Jono Alderson

Or your competitors might mess something up and you'll never know.

Robert Hansen

You'll never know?

Jono Alderson

No. If you are doing good stuff and you are investing in improving the quality of your website over time, direction that you will go up. And that there's no rocket science there. There's no magic that happens and we see evidence of that time and time again. But yeah, you'll never isolate what is...

Robert Hansen

What is so the SEO of old is really just a collection of strategies they have to be implemented in uniformly across your stuff. And it had better be working with really good content to begin with, otherwise forget it. That's pretty much it.

Really good lead funnels. Which I think is interesting because that means that a lot of the old school kind of hacker sort of short-term wind-type stuff, it's almost kind of counterproductive.

Jono Alderson

It still exists. I'm firmly in the school of brand building. I think that's more interesting. I think just personally, my preference and the world I want to help shape and live in is one populated by brands that are meaningful and helpful, and stand for something.

The alternative to that is you can fire up a 100 domains, spin up a whole bunch of cheap content for Viagra and casinos rank really well for two or three days based on some tactical manipulation and then burn everything. And you can rinse and repeat that to a degree.

That does get far harder and far more expensive. And the success rate is far rarer. It is still viable. There are still lots of people who do that. Some very successful, very happy people. I think that's less interesting. And I think that that will continue.

Robert Hansen

It's a harder road ultimately.

Jono Alderson

Yeah. And for a long time, it's been far harder to do it the evil way than the good way.

Robert Hansen

Yeah. Which is usually it's the other way around. So, we'll talk about more evil stuff in a minute, but you kind of mentioned that there's this thing that's going on in the marketing industry and it's been going on for a while. I've been watching it basically since the beginning.

So, I was employee 11 at ValueClick many, many moons ago. So, one of the reasons I decided to leave that company was because I sat right next to the sales team. And they very friendly ladies, and they would make these phone calls like, "How much inventory do you want to buy?"

And they would tell them you should have really beautiful ads. Make it really engaging, really beautiful ads. And then, I was sitting there thinking about it, and I'm like, actually, if it's like Nike or whatever, they should have the ugliest ad you could possibly imagine.

Because that means fewer people are going to click through. And ValueClicks model was a Pay Per Click model. And from your brand's perspective, you don't...

Jono Alderson

All the brand exposure, but no clicks.

Robert Hansen

Right. All the brand exposure, no clicks. And I'm like, they're literally giving the exact wrong advice. And that's sort of had this tingling, and I'm a pretty ethical person, and I'm like, "I just don't think that's going to work for me."

I can see the writing on the wall for me and once I start feeling like I am contributing to the problem, it's a much harder thing for me to stomach. And eventually they got bought by DoubleClick and DoubleClick got partially by Google.

And so, a lot of my old code is now inside Google. So, I actually have a pretty good insight in how that all works. And then, the other thing is, I was in part of another company as advisor, a first company I ever advised a company called Adometry.

Their whole business model was consuming logs from companies doing advertising, analyze it and see how much fraud they thought was going through the system.

And it turns out it's an enormous amount of fraud, like an enormous amount. And they were eventually bought by Google and shut down effectively. Which is fairly telling about where Google is.

Jono Alderson

As you say, answers that product. And we all know anecdotally that a huge number of the clicks on both sides, both on the kind of side and people's inventory and going through Google, we know that a huge amount of that is bots.

But yeah, it's a very tactical line that Google has to walk, how much they want to police that, how much they want to talk about it, how much they want to admit. This is a bigger problem than Google buy. There are a lot of bots on the web and it is in various people's commercial interests to cheat this.

Robert Hansen

There's fellow I met who was willing to give me his logs and he ran a fairly large ad network. I was trying to do some analysis on the likelihood that someone was ever going to click on, do not track within different browsers. Extremely rare thing for anyone to actually go through the process of doing it.

Jono Alderson

Extremely frustrated. They deprecated because it would've been quite a handy tool.

Robert Hansen

Would've been great. But they knew full well it wasn't going to work.

Jono Alderson

Far too cumbersome.

Robert Hansen

Because there was no good controls and it doesn't do anything.

Jono Alderson

Nobody respected it.

Robert Hansen

Nobody respected it.

Jono Alderson

I might be the only human who's ever implemented it in Google Tech Manager to control analytic. Like nobody bothers.

Robert Hansen

But the reason I thought it was interesting is in some browsers you had to click like three times to turn it on. In Google you had to click like seven times to turn it on.

So, they're intentionally making it very difficult. And so, there had been some research paper long ago that basically said it was like order of magnitude less for every click you had to go down to turn on or off every feature.

So, effectively from a UI perspective, you should hide any features that are very unlikely that anyone's ever going to need. So, I was going to do the analysis to see if that's in fact what happened. If the companies that made it more prominent or more likely to see it, then the companies made it less prominent.

So anyway, I got access to this guy's logs and I was dumping this one header, the do not track header and the user agents to figure out which ones were correlated. And for whatever reason Internet Explorer was off the charts. It was like 10 plus percent Interesting. Which is impossible.

I mean, it's still, I forget what it was now. Like five clicks or something. Literally impossible. There's just no way millions of people have gone down those five clicks. Zero chance that that is true. And I'm just looking at these logs, I'm like, "This can't be right. It can't be right. There's just no way this is true."

And I've suddenly dawn on me, that's just how many robots that I just detected.

Jono Alderson

Of course they all use IE.

Robert Hansen

Right. They use IE because they just want to blend in and pretend these are robots and they just grab a user agent and that's the one they're using. And that was just one robot that I detected. Now imagine how many more as a percentage, I didn't detect using that one technique.

And it's probably, I'm going to guess here, but 20%, 50% of the traffic that we're seeing is robotic. And now you extrapolate that to Google and conversations that they've had, they're like, "Well, Adometry says that we only have about 5%, 10% fraud."

And one of the mega companies involved basically said, "Are you kidding me? Because that means that my traffic represents all the fraud on all of the internet. I think you are wildly misunderstanding how much traffic is fraud."

So, the heart of my question, that's just all background. And my question is, it seems like there's some sort of moral bankruptcy going on here. Everyone knows that this is messed up, who's bothered to measure it? Including the people who run these things, including people like me who used to work there. What's going on? Is this purely just too much money to say no at this point?

Jono Alderson

Yeah. But also I don't think we have a better option. I think we are tied to the most of the commercial internet being based on either on both sides. Most of the commercial internet being tied to per click advertising mechanism.

So, whether that's Google ads directly in there, or banners on people's sites. And no matter how hard we try, we can't seem to find an alternative to that. And the challenge of that model is one of scale and management and micromanagement.

Of course you end up with intermediaries and ad networks and fourth parties and exchanges and ads being handled off to different systems.

And then, that is all invisible because no brand, no company, no marketing manager is ever going to go with brand and manually establish a, manage and maintain relationships with different websites, with different platforms, vet all of the individual domains their ads are going to show on.

It's a numbers game. I want this banner in front of 10 million people in the hopes that I get 1% of them to click through. And it's wildly inefficient. And it is still a wild west. It's cowboys and it's almost completely invisible.

Nobody has come up with a better option. Micropayments as a mechanism has failed. Subscriptions as a mechanism has failed. Direct partnerships doesn't scale and has failed. We don't have an alternative for how the commercial web should work. And still nobody comes up with the one so it just gets bigger and bigger and bigger.

And then, SEO's getting harder and harder. So, there's more money being plowed into that. And analytics is getting harder and harder. So, there's more money getting plowed into it. It gets more and more anonymous.

Browsers themselves now baking in more privacy controls like Firefox stripping out Facebook's tracking URLs. It's going to put more money into ads because it's harder and harder to prove that stuff's working.

So, you go broader, now I need my advert in front of 10 million people in the hopes that I get knocked 0.1%. And it just becomes more and more homogenous and these systems get bigger and bigger. We, we don't have another way of doing it. It's horrible.

Robert Hansen

It is. So, let's talk about the Google antitrust lawsuits. So, you and I know a number of people who've been asked to go present against Google in these various different cases. They've lost some cases, like in the European Union.

So, if anyone's curious about Google's ethics, it's proven in court. This is not up for debate at this point. What do you think is going on? What do you think Google is doing that's anti-competitive?

Jono Alderson

I think there is definitely a strong argument that they prevent innovation from potential competitors. There is no way that today, anybody new will enter the maps space or the flight space. There's no point you can't compete with Google's products.

And the crux of the argument around that is, are they using their monopoly or monopoly-like scale to unfairly promote their own properties in place of where a competitor might be? That's part of the challenge.

If I build a flight comparison site, Google will continue to promote their Google flights in the first spot. It's a really great service as a user. I love it. But I kind of like the idea that somebody else might be able to try and build something better or at least different.

And that market's now completely closed off. And there are 100 variations of that story. There are spaces where you just can't compete.

Robert Hansen

It's the great scrape and replace.

Jono Alderson

So, there's that as well. So, also there are spaces where I know that if I start something, they will just steam me because I will build a proof of concept that Google will look at and say, "Okay, we can do this better."

If Google flights didn't exist today and I built Hipmunk, which is exactly what happened, Google can just look and say, "Okay, this is a space that we can work in. We'll just roll out our own thing."

And we're in a position where I want to build a healthcare startup. I want to build an eBag delivery service. I want to build a dentist on demand service. Tomorrow, there is a reasonable chance that Google will launch its own version.

I can't take any story to a VC and say, "I'm going to build X." And they're going to say, "Have you considered that Google's a threat?" There is no answer to that. There are a universal threat against any sector. I'm really conflicted because as an end user, this is great for me.

And I think there are some phenomenally powerful and positive network effects. So, Google are in the driverless car space through... What's the name of that company?

Robert Hansen

I think they just sold that though.

Jono Alderson

Oh, okay. Well, let's pretend they didn't for a second. So, they're in the driverless car space and let's assume that they win because they're Google and they'll leverage their monopoly like status to crowd out the competitors.

You've got a whole bunch of driverless cars driving around urban spaces that also happen to be acting now as Google's mapping system, which also then happened to increase the accuracy of Google Maps, listing businesses opening hours.

If I was going to go to the bakery to get a class on and Google now knows that they're closed, that is an incredible network effect that is so really, really useful as an end user.

But the amount of space that closes out and prevents businesses competing across multiple sectors and just reinforces their monopoly power is really, really scary. And I think that's a big part of what the various EU and US efforts are going to try and do, I try and break those entities up.

I don't think that even helps. I don't know what that means. I think breaking it up is technically is really, really complex. And even if these businesses are different entities that don't share data, they're so big and they're so unanimous and it's not really going to help.

And then, to your point about the scraping the place, this is now hitting the content side, which isn't just flights or different verticals. This is all content on the web.

So, as a content producer, writing content, in order to try and help Google's users in Google and solve these problems and have all these great dentistry guides, I produce in-depth articles that I research that I employ experts for, that I produce media for that I keep up to date, that I really invest time and energy and money in to.

I hope that I rank and I hope that I get a couple of clicks, but maybe not, maybe I get pushed out by Google's own dental service. But let's say I'm fighting to survive at the bottom, and then the next thing that happens is Google scrapes that content from my page and they just show it directly in the search results.

So, that a user who's searching, "Hey, I've got toothache" just sees a paragraph at the top of the page that summarizes my text. Maybe it doesn't even replace it in the bottom. Maybe it just summarizes it. Maybe I get a little attribution link in a very faint gray color. Maybe, maybe not.

And now, users great problem solved. "Google's amazing. They've solved all my problems." As a content producer and a business. I have lost my commercial incentive to produce the content that Google used. And there's some,

Robert Hansen

And they used to penalize people for doing exactly this.

Jono Alderson

Yeah. They have documentation saying, "Don't do this." phenomenal. That's a rabbit hole. So, this is really tricky. And there isn't an answer to this. I think the SEO industry is focusing in the wrong place at the moment.

All of the focus we had this morning in some of the sessions, all the focus at the moment is on this concept of zero click searches. Which is, there are various studies that show there is an increasing proportion of people who go to Google type something and then don't click anything.

Which is the inference is that they saw something like the rich snippet, the paragraph at the top, and they didn't need to visit the underlying sign.

Robert Hansen

They were trying to do math or something. They got the answer.

Jono Alderson

All the tooth package.

Robert Hansen

They don't need to go to calculator.com or whatever to do the Math.

Jono Alderson

Google can provide a frictionless user experience and they can eat up whatever business model. This is more and more of search and this is a direction of travel. This isn't done. More and more of the experiences that Google want to offer our solutions in Google not to ship you off to some other website.

And the SEO industry, because it's always been French and because it's always been tactical and because quality isn't owned by anybody, is built, measured and managed quite rudimentary.

It's all based on clicks. How many clicks did we get from Google based on our rankings and how do they convert? But how many clicks do we get? And suddenly clicks aren't a thing anymore. And the entire industry is focused on how do we get our clicks back? How do we get our users back?

They were never your users. They were Google’s, we just kind of rented them for a bit. Whilst it was easier for Google to do that, nobody is talking about how do we build business models which still thrive in a world where we're not reliant on getting clicks from Google.

It's hard. I don't know what the answer is. I've spent a lot of time thinking about this. Again, micropayments have failed, subscriptions have failed. I don't think there is necessarily an answer, but that's the conversation that I think we need to be having. Because this is only got it.

Even if Google gets broken up, even if all these lawsuits happen and they start to pull apart things like Google flights, there is still an underlying problem that Google don't really necessarily want to send users to other websites. It's not in their best interest.

And there's a conflict right at the heart of the model that's not going to go away regardless of what happens legally.

Robert Hansen

So, that's perfect segue. So, I talked a little bit with Elijah on very last episode. So the very top, especially more contentious things like hotels in Atlanta or whatever, you are not even going to see a single search result for the first half of the page.

Jono Alderson

I don't know if that's necessarily a problem. I think there are many types of searches where the best type of result is not going to be a link to a webpage. It's going to be a map and some options and a calculator and a tool.

But yeah, from a conventional SEO standpoint, where I used to rank first at the top of the page, I might still rank first, but I might be 30 miles down the screen.

Robert Hansen

Yeah. I mean, you're so far down, literally no one will even see it.

Jono Alderson

Yeah. And that also comes with some really interesting challenges around reporting and accountability and strategy because all of my forecasting no longer makes sense. The concepts we share in the language we use no longer make sense.

If I'm saying to my C-suite, "Yes. We rank first for this." That's a very different understanding of what that means. The tools themselves...

Robert Hansen

But you rank first?

Jono Alderson

Yeah. Great. But nobody's ever seen us the tools themselves all have different models for this. So, things like Google Search Console, which is their own model for collecting, giving you some reporting counts, individual listings in the horizontal carousels.

So, I might rank sixth, but that might mean that I'm an off-screen panel in a carousel that I have to scroll across to get to. So, none of this makes any sense anymore. Which funnels loads more money into adverts and Ad Words, which maybe isn't a coincidence.

Robert Hansen

It seems like it's the measurement should either be just visibility or not. We have it or we don't or maybe like a number of pixels.

Jono Alderson

Something like commentures like we have in traditional, you own 13% of this page. Great.

Robert Hansen

Yeah. And that makes a lot of sense. That's a great way to measure it. But what I think was happening is the SEOs are just not even on the page anymore. They're below the fold. You have to scroll to get to the content.

Jono Alderson

There's a deeper challenge there as well, which I'm sure we're here about at the conference. There's a few people in this space around what does that ranking even mean?

Because we are sat three feet apart, we probably get different rankings. We have different search histories. I'm logged in on my phone and different devices, et cetera. And a whole bunch of other factors.

Robert Hansen

If you're in private mode or not in private mode.

Jono Alderson

And which country I'm in, which my browser language settings, you're probably enUS, I am enGB if I remember to change it. All of these are going to affect individual rankings for individual keywords.

Over time, it's going to change whether I've searched before. All of this is fluid. It's really, really hard to make any of this accountable. And then Google sweep in extract all your content, all of your clicks disappear.

And I'm sat here saying, SEO strategy is all about investing in quality and making your business better. How on earth do you build that business case.

Robert Hansen

How do you do that? Because now you're so far down the fold. I'm just getting the sensation that Google just won where you have to have SCM, you have to buy ads or you're just done. I mean, it's a very weird world. We've just entered the brave new world here.

Jono Alderson

It is. And I think the broader pattern, the way I like to see this is what's happening is that there are many sectors and many types of queries which are solved. You said calculators. Now, there is no incentive for anybody to build a website that helps you with basic math problems now, because Google will solve the calculator thing in line.

There are more and more types of searches where the problem is solved. And suddenly we've gone from a paradigm where Google really wanted and needed websites to produce content to fill its inventory. Inventory is now full.

There is no need for anybody to ever write a lasagna recipe ever again. They have every permutation and they have enough which are as close to perfect as to make near damn no difference and the world really hasn't noticed.

And we're still all trying to compete and churn and put our versions of all this stuff into the space. These are closed systems now. There is no way to get in. And any of these spaces where the ranking first is all the way down here.

Having better content isn't going to help. Trying to do conventional SEO isn't going to help. You have to change your business model. You have to move your location so that you are in a better space on the map.

You have to have a licensing fee to get your information fed into Google flights. These are different ecosystems. And really the whole game has changed, definitely.

Robert Hansen

All right. So, we agree on basically every single thing we just talked about. Now, here's where I think we might start disagreeing.

Jono Alderson

Contentious toetry. AMP.

Robert Hansen

AMP. This is the one area I ever remember having an argument with you about.

Jono Alderson

Yeah. I'm the pariah. The whole SEO industry hates and I remain convinced that it is/was, maybe we'll come onto that.

Robert Hansen

All right. That sounds interesting. So, before we get that, let's describe what it is though, because I ask people all the time, have you heard of this? And everyone universally says no.

Jono Alderson

One of the things Google consistently does very, very badly is marketing and product launches.

Robert Hansen

That's so funny. That they're in that business.

Jono Alderson

So yeah, Google Wave, Google search. The AMP is the latest. So let's start at the beginning. Most websites are bad and most websites are slow. And there's a whole bunch of reasons for that around education and resourcing and prioritization and business models.

But largely speaking, that's the general state of things is pretty poor. The mobile web in particular is very bad. Because there have been various ways to build mobile sites and various best practices and standards over the years.

And most websites have got caught up somewhere between two or three of those iterations. And they have mobile sub domains and sites that don't load properly on different devices.

And in particular the mobile web is extremely slow and even slower and worse in emerging markets and developing countries where there is lots of potential ad money to be spent coincidence.

So, Google's solution to this, was initially to spend 10 years to tell us all it was the year of the mobile. And we should sort ourselves out and build better sites, which just didn't happen. There is not enough education, there's not enough incentive, there's not enough motivations for them for whatever reasons.

Robert Hansen

And it takes time.

Jono Alderson

This stuff is hard. And this is a key part of the problem. So, the solution was to launch a product, an idea called AMP, which was initially Accelerated Mobile Pages. And the premise was what if we could give you an opinionated subset of HTML, which was inherently always fast.

And a big part of how it would do that would be by preventing you from doing 'bad things'. So, a large part of why the mobile web in particular is slow, is because this page loads five megabytes of JavaScript and these images are huge and uncompressed. And you are doing a whole bunch of weird things with cookie pop ups.

Robert Hansen

Like in frames.

Jono Alderson

Frames are fun. And you've got ad networks pulling ad networks, pulling ad networks. So, the premise of AMP was what if you could build a lightweight version of your pages? We make it super easy to build that We codify what good looks like. We are very opinionated on what bad looks like, and we don't let you do any of the bad things. And rather than have a crappy, shitty slow thing, you have a great thing.

But it was fought with challenges both technically and politically. So, it was extremely limited when they launched it. It was very hard to brand things with it. And you had to build an entirely separate version of your pages in an AMP template.

And if you're saying the part of the challenge of the reason that the mobile web is slow is this stuff's hard and complicated requiring people to build another templating system, probably not the most sensible approach.

And they also did what they do badly with every product release was they do it iteratively. They start off with a product, which is so far sub-MVP. Everyone gets very excited about the concept, gets very annoyed about the actuality and then abandons it. Which is exactly what happened here.

So, the initial versions of it were far too early. You were very limited in what you could do with branding. There were some very, very major limitations in how you could use and deploy adverts, which obviously is quite a key component of how many sites monetize.

And there were some weird technical decisions which looked on the surface like political decisions, but I'm pretty convinced weren't like all of your pages would be cashed in Google's own cash.

Robert Hansen

I feel like that was intentional.

Jono Alderson

So, it was partly a technical strength due to how cost domain... Obviously your area of expertise. How cost domain requests to management work. That there was no mechanism by which they could cash your content and then serve it on the right domain.

So, they invented a whole new concept, which many people come across now have signed exchanges. A lot of the new emerging stuff around performance has come out of AMP. Come onto that in a minute, maybe.

Yeah, you had to have all of your content served from Google's cash, and yet simultaneously the controls from managing that were horrendously put again, they just launched a bad product too early. So, there was no way to purge the cash.

So, if you were cash in e-commerce pages and your product went out stock, your price changed, or you had some legal requirements, you can't touch Google stuff. And there's a whole bunch of other mess around that. But essentially they launched it badly, technically and politically.

Robert Hansen

And you inject the wrong character and the whole thing breaks.

Jono Alderson

Yeah. It was super, super sensitive. So yeah, on one hand, they're building chromium, which is phenomenally forgiving. I can put the world's most invalidates HTML in, create a soup and it will do a reasonable job of working out what I meant.

AMP, you go, "Okay, somebody has accidentally put an ampersand in the wrong place," And you just get an arrow page and there's no way to recover from that. And then, you also needed to manage the relationships between your pages and your AMP pages in the same way you might do with other SEO stuff like internationalization. Everyone got it wrong.

Everyone did a bad job. You also had to load the AMP JavaScript library across domain from Google's data centers, which is an interesting political and legal requirement.

Again, all of this was just bad product design. So, I still quite firmly believe it is a good product or had had good intents to be a good product. I don't think people can make and maintain fast, high quality websites. I think the motivations aren't right.

Robert Hansen

Okay. So, here's where we started diverging, because up to this point, I agree. But there was an assumption that they're going to do a better job because they were actually de-ranking people who had potentially even faster, better working sites.

Jono Alderson

There's a whole bunch of quasi legal stuff around this written by a bunch of people in Texas who had a vendetta against Google. And I've been through all of this. And a lot of it's furious. There's some interesting stuff in that.

So, Google for trending news topics in particular and a bunch of other areas, Google has a top stories carousel, which is the horizontal swipey thing for emerging news. And if you wanted to be in that, your page had to be on AMP.

Which was widely perceived as a tax on the news industry. And there is nobody who hates doing what Google tell them more than the news industry. They already have a very contentious relationship with Google. None of those people are friends. And they see Google as the enemy.

Robert Hansen

They're right to think that.

Jono Alderson

Maybe.

Robert Hansen

I think they are.

Jono Alderson

Perhaps. I think they would prefer a world where Google is subservient to them and they dislike that that's not the case.

Robert Hansen

Yeah. Agree.

Jono Alderson

So yeah, if you wanted to be in that, you had to be on AMP and there's no two ways around that. The intentions behind that were never to create that as a barrier to entry. It was that the mechanisms for determining what a fast high quality page were did not exist. And it's only as a result of the AMP project and the evolution of the core vitals metrics came out that's now universal and that requirement's gone away.

It was a really bad product decision to roll that out before that ecosystem had evolving. Again, so much of this is just Google being really shit at product. And all of what looks like political maneuvering and coincidental evil just feels like bad product management on their side. It's really quite shocking.

I mean, it's hard.

Robert Hansen

Well, there's other things along those lines. Like you could only use Google Ads, for instance, you can use their competitors ads.

Jono Alderson

Yes, again, just bad product because they launched it too soon and since they've added support for others, because core premise of this is you can't load things in a way that's slow. Google Ads was quite fast to be okay, now we're compatible.

But it turns out that all these other networks need to fundamentally rewrite how they work in order to be fast. So yeah, you can only use the fast ads and it just so happens that Google is the only one who has a marketplace for the fast ads.

Robert Hansen

That's surprising, isn't it?

Jono Alderson

Yeah.

Robert Hansen

I feel like that's more than a coincidence because there's a lot of coincidences here.

Jono Alderson

There are. But I still think this stuff is good or at least this stuff can be useful.

Robert Hansen

One other quick point, I want you to finish. I just want to make sure we get all the bad stuff out. So, how to see this, if you want to do this at home is go to desktop search for CNN and Trump, “CNN, Trump”. Then do the exact same search on mobile.

What you're going to see is a carousel, this little swipey thing. And you might see something about Trump on CNN and you might see another thing, and then you're going to see Breitbart and then you're going to see NBC or whatever.

So, you specifically said where you wanted to go and what you wanted to see, and now Google will serve you up something completely different in the exact opposite direction.

Jono Alderson

Yeah. There was a really weird moment where they started allowing you to swipe between stories from within it. That

was nefarious. From what I can tell, there's a whole bunch of competing interests.

Robert Hansen

That’s just devaluing the entire news industry completely.

Jono Alderson

Absolutely. I wonder if that was an early attempt to do something like Google Discover before that got enough traction. The idea that we want to keep you within the Google's ecosystem. Swipe through. We know this was a terrible plan. Really, really bad. The other really evil thing that they accidentally did was, one of the key advantages of adopting AMP and being compliant was they can preload and pre-render your site.

As I click on my search result, it loads instantly. Or at least appears to because it's pre-loaded in the background. They could only do that because of the way the AMP worked and because of the way the caching system worked. Specifically the relationships between the different domains. Again, this looks evil. This looks like an advantage. You're exclusively giving, above other sites, to the people who happen to use your preferred tech stack.

All of these questions come down to, yeah, this is a product. The way this thing works, the only way it can work is that unfair. All of the news industries gripes against this thing are, you unfairly rewarded your own technology over all of us who are doing the bad things. Please let us do the bad things and also get the rewards. Tricky.

Robert Hansen

This also reminds me a lot of how they treat Chrome. It's like Chrome is the thing anyone can have. There it is. It's open source. There you go have it. It's Chromium. Just download it. You can do whatever you want. But they won't give you the most critical components. They won't give you the ability to update it. It might be useful if you're launching a browser. You're thinking, “Oh, that’s okay. That's a relatively easy thing to build.”

Well, it isn't though. Because now you got to somehow get all of the bugs that they're producing. You got to correlate them. Test all of them. They're already six months ahead of you to begin with because they have products that they're launching. You want to be in parity with their product.

Now you're getting these bugs. You're having to backfill them. You're never going to be anywhere near where they are. Yes, it's technically open source. But you're never ever, ever going to be able to use it.

Jono Alderson

We have the same problem with WordPress. That yeah, it's open source. If you don't like the way the leadership and the politics and the marketing work, go and fuck it. Build your own version. Except, try and build a community and maintain parity and keep it. It is just not viable. Yeah, it's open source, but it's too big to fail.

Robert Hansen

Why do you like AMP? What's great about it?

Jono Alderson

It makes fast sites in a way that I don't think practically people can do in the world. I can handcraft a site from scratch. It will be lightning bloody fast without using AMP. But real businesses in the real world just can't. The bigger they get and the more complexity they add, that just becomes impossible.

Robert Hansen

Why not just have Google say, “I want a fast loading site.” That's it. If it is speed, and it can't scroll off to the left or do something stupid on mobile.

Jono Alderson

They already said that. They have for decades told this around that. What AMP does different, AMP is now. It comes with a couple of iterations. It is essentially just a JavaScript library. It's not different from react or jQuery. The thing that makes it different is opinion. That it says, you cannot load a video and have it autoplay. You have to do lazy loading and do a facade cover over. A whole bunch of other rules.

It is only those constraints and enforcing them that enables people to build fast sites. Because they don't know what they don't know. I did a conference talk last summer with 120 slides of how to put an image tag on a web page. Because it turns out that even the most simple, ‘How do I build a reasonably not shit website?’ there is a phenomenal amount of depth and complexity in the modern web. People just don't know.

Most developers and most companies will never get past slow because it gets harder and harder. There are more moving parts. They're all running 20 different ad networks, autoloading video, images are complex and fonts are slow. People can't get it right. It is only by enforcing a set of constraints that they can. Then the constraints are evil. Because who are you to tell me that I can't autoload a video on my site. Well if you can, autoload it. But it won't be fast and it won't be compliant. You won't be eligible for the pre-caching because of the way it works.

I think a lot of the bad stuff has been fixed. You can now self-host. You can now opt out of the cache. The repository itself is owned by the OpenJS Foundation. They're going to move the files off Google servers. There's all the components so readily available independently. It's a much better thing. But they did such a bad job with the product launch and the marketing and the management of it. The name is utterly toxic. They tried to rebrand. They launched Bento last year. Which was, ‘What if we give you just the components?’ You want to put a carousel or an image on, just take this. Use our JavaScript library. We'll pretend it's not AMP even though it's the same thing.

Again, so badly launched and so badly marketed that it just failed. It makes me really sad. I got involved quite early on as part of the advisory committee in an attempt to work out what the hell was going on in all of this and to maybe stare at it a little bit. That's dried up. It's gone really quiet. I think a lot of the key people who were working on it at Google have left as part of a bigger exodus. With the whole, all the performance people at Google have gone Shopify. Quite interesting.

It's just now essentially abandoned ware. Hopefully, it'll get a kick in initiative once it moves properly into the OpenJS Foundation. Don't know. I think it'll get quietly retired. A lot of what it's done is now just part of Chrome. Code vitals. The signed exchanges stuff is really, really cool. The signed exchanges stuff solved all of the pre-caching, pre-loading, cross domain stuff that you had to be on AMP for. Now you can do that yourself. It's inordinately complex. But you can.

I don't know if it's needed anymore. Some of the early strategic discussions we had in the group were, if this is toxic and it's too late salvage. Maybe we can reinvent it a bit to be pro-nightly. Maybe we can make this an experimental tool. That edge is all the new cutting edge evolving tech. It was that for a while. Now maybe it's done its job.

Robert Hansen

Or just get it installed somehow into WordPress. Now you got 43% of the web.

Jono Alderson

Yeah. Just so happens that some of the other people on the advisory committee happen to be part of Automatic. We did look at that. WordPress has just moved to a block based system of building pages. Blocks, components, Bento. Maybe this is all a bit too complex.

But I think if I were a small to medium sized business who didn't have expert developers and who wanted to actively compete in SEO, didn't want to spend 200,000 euros building most of the static brochure site, AMP would be absolutely the right way to go still.

Robert Hansen

Speaking of more drama on the Google side. Let's talk about negative SEO. One of my favorites. Yours too. I know we've had long conversations about this. We couldn’t stop arguing. As recently as this trip, I was talking to someone, a former Google employee. He told me that he indeed tried to start measuring negative SEO internally. What that might look like and who created custom reports. It got zero traction internally. I'm shocked to hear that. What negative SEO is for those who are not initiated? Would you please?

Jono Alderson

We've talked about links as a voting mechanism. If one site links to another, that is a vote of endorsement and AdScale that hugely influences where my site ranks. It arguably might be the most important fact. The quality and quantity, which is an interesting word, of your link profile is going to significantly influence where you rank versus your competitors.

We know that Google counts and looks at and evaluates the strength of a link based on factors like how popular is this domain? We also know that they look at the broader context of a link. Like what is the text that is linked? What is the text around that? What is that page about? How reputable is this site? Those things definitely have an impact on your performance. Therefore, the inverse is potentially true.

Google insist there is no such thing as SEO even when they acknowledged that there might be. They say almost nobody needs to worry about this. However, they provide or at least have historically provided a tool in Google Search Console which lets you explicitly disavow links from other websites. Which would imply that if certain types of other websites do link to you that could be problematic and might be the thing you want to manage. If say 10,000 Viagra sites started linking to my dentist site tomorrow, I definitely don't want to be ranking for Viagra.

Yet we see anecdotally evidence that this thing could, maybe does, maybe, potentially impact rankings. Quite often, it's much more nuanced than just, I will throw 1000s of spam links at your site. Quite often it's, if we know that Google is looking at context and they're trying to evaluate what pages and websites are about, maybe I can subtly shift that in a little bit of a direction. Maybe we can make it think that your cat blog is about dogs. Not quite to that extreme. But you can steer. Yet they insist it's not a thing. It definitely is.

Robert Hansen

Well just taking it back a step. Let's say I go over to some business. I literally drop a bomb on it. It just destroys. The data is all gone. They're serving on a local machine. Google can no longer find it because it is literally in flames. They might drop out of the search engine pretty quick. If they were above me and they're no longer there, where do I go?

Jono Alderson

It is definitely in your interest to go and bomb all your competitors.

Robert Hansen

Yes, I highly recommend bombing all your competitors. If that's possible, it might be possible that there's less crazy things you could do that might have similar effects.

Jono Alderson

Conventionally, we're talking about, can I send some nasty links to somebody? But if you zoom out a bit. You start to think about manipulating businesses, people, social engineering, there are definitely things you can do to downrank your competitors, or up on cue. Yeah. It's a pretty nasty space.

Robert Hansen

It is incredibly nasty. They really, truly seem to, some like this engineer that used to work there, obviously know it's real. They're here. It's like trying to disavow error. I can't see it so it's not there. What are you talking about?

But we know that they're not even looking anymore. How we know this is, there was a letter written to another SEO who used to come to this conference. Where he basically said, “Hey, I believe there's some negative SEO going on this thing. There's nothing else wrong with the domain that I can find. Can you please take a look and see if there's any anything you can see?”

One of the head quality guys over Google, he said, “I've checked the systems. You don't have a manual penalty. You don't have a programmatic penalty. It's probably a problem with your site. I recommend going and checking these five things.” Then below that there was an up arrow that said, ‘This is the block of text I send to people when I have in the hundreds of times checked on this thing. I'm no longer doing this. Just go check it yourself.’

There are tools. The tools are not set up to work properly. They do not find any negative SEO by his own determination. Therefore, he doesn't bother to check them at all.

Jono Alderson

Yeah. There's another really interesting interaction here. Which is that we know that there are scenarios and complexities where Google gets things wrong in the interplay between canonical URL tags and different types of redirects with different HTTP statuses, different multidomain stuff and too many hops in the wrong status. We know they make mistakes.

We know that there are scenarios like, if I have the wrong redirect pointing to the wrong domain with the wrong thing, they will index the wrong page. Or they will index the white page. They'll put the wrong URL on this. We see all these. Certainly a scale for bigger or complex websites. Knowing that, yeah, maybe I can tactically point this here in the right way.

Google will confuse something, invert it. Look now my brand name is in your listing. There are definitely lots of little edge cases where you can play with this. They shot them quite quickly when they see them. They definitely have a whack-a-mole capability for a given site or a given domain where they can go in and tweak this. But yeah, they continue to say this just isn't a thing.

Robert Hansen

My theory is they truly don't know what to do about it.

Jono Alderson

Yeah. It's so abstract.

Robert Hansen

It's so impossible. There's so many different ways to do negative SEO. Negative SEO is to Google as Google is to SEO. In that they are so overwhelmed by the possibilities of negative SEO. From their perspective, it is insurmountable problem. From the people I can reach who actually have a clue about what's going on. Whereas SEO is so insurmountable. What are they going to do? Just try to do your best and hope it all works out.

But I think negative SEO unfortunately, is that freight train that is coming. That the more SEOs who figured out that this is the thing they can do. Yes, I can't do traditional SEO. I can't.

Jono Alderson

Yeah. It's hard to rank higher. But maybe it's easier to knock everybody else.

Robert Hansen

Yeah. That's scary.

Jono Alderson

Traditionally, SEO falls into either, I'm black hat and I'm breaking Google's rules and I'm manipulating stuff. Or I'm white hat and I'm one of the good guys. The next step is into kind of, now I'm doing stuff that's illegal.

Even a lot of the black hat stuff might be unethical, but it's rarely illegal. But when you start getting into, now can I socially engineer my competitors to make sure that they fail to renew their domain so that I can drop cache it, we start to cross interesting lines. That definitely happens in the real world.

Robert Hansen

How would you say Google is viewed amongst your peers? I remember asking this question many years ago, maybe a decade or more ago, to some of the SEOs who used to come. Their opinion was, if you could get rid of Google today, you would in favor of Bing or whatever. I don't even think Bing existed back then. But what do you think their opinion is now? Is this a love-hate relationship? What is the stat? What is the mood amongst your peers?

Jono Alderson

I struggle with this because the SEO industry itself is fragmented somewhat. I think now, again it was all, let's find some tactical thinking opportunities. There are now two, maybe three distinct types of SEO. There are the tech SEOs who are very interested in XML, Sitemaps, and canonicals and joining up a lot of that. Maybe some of those are good. Maybe some are bad. Maybe some of them are doing manipulative dodgy stuff.

There is the content people who are like, we write blog posts. We do informative guides, etc. Increasingly, the largest part of the industry is now just digital PR. Which is, we want to get links. But we want to do it in a way that doesn't look like we're getting links. We are going to produce a spurious piece of research and hand a bunch of journalists to try and get coverage for it. It will look like authoritative resources. Like, “The Guardian are linking to us because we're a great brand.” But it actually is because we did some tedious research.

I think they would all give you a different answer. I think the digital PR scene certainly see Google as a very lucrative resource. There is a phenomenal amount of money in agencies and to a lesser degree independents, producing that research and collateral for the clients who would otherwise struggle to rank. I think this is an interesting fusion of the SEM versus SEO. This is candy sugar rush spiky SEO in a way that's never really existed before.

I can go, I'm going to spend 20,000 euros to produce some infographic key interactive researchy thing that I will have a reasonable chance of get picked up by a bunch of newsletters. My traffic will spike for three weeks. Then I want another one. They are loving it. There is a huge explosion of the PR space that very much see Google as a cash cow. I think everyone else still sees them as a bit of an enemy. They are not. They're playing better than they ever have before. Their documentation and their interactions get better.

But they still see us as the enemy. I don't think they like SEO. I don't think they really want to acknowledge it and acknowledge that they are relying on it despite increasingly becoming so. Again to the point that the index is now increasingly complete. They don't need us as much. They're starting to go cold again. I think it's a bizarre relationship where we've never been friends but we mutually need each other. That's changing again. It would be interesting to see where it goes.

Robert Hansen

You guys have crazy good access to Google as a result of what you guys are doing. Not just the AMP stuff. But the SEO engine which is exactly what Google needs for 43% of the web. You've got to have really crazy good access to these Google employees. Where you can maybe even help shape the future of the web. Not just consume whatever technical things that they're producing.

Jono Alderson

Yeah. This is fascinating. Having transitioned from an SEO who worked for businesses and in agencies on the receiving side of Google's guidelines, to now being in a position where the relationships are still weird. But there are scenarios where we end up shaping what they do, how they react to and prioritize things. Which is pretty weird.

I think that said, the thing I have to keep reminding myself and I'm really learning to be true is, there isn't such thing as Google. There is no company called Google. In the other direction, there are just teams working on projects. This is a big part of why they are so bad at marketing and product. They are utterly fragmented internally. Yes, there are overarching strategies and directions and big products. But all the evidence I see suggests that there are lots of itty bitty projects. It's almost like a university campus on one part of the culture.

But yeah, we interact with loads of those different little teams. It's always a bit tense because they really don't like interacting with the outside world. It's really alien for them. It's really awkward and difficult for them to reach out. There are certainly scenarios where they're either not meant to or not allowed to for various legal reasons, and they have policies within policies. It’s messy.

That said, the Google teams within the WordPress ecosystem are very open. I find that really interesting. There are six or seven of them working mostly across performance in the WordPress ecosystem. But also bits of security and other stuff. They are some of the smartest developers I've ever encountered. Incidentally, some of them worked on the AMP plugin for WordPress which I run on my site.

I love this. I used AMP to build my own personal website powered on WordPress using the AMP plugin. Which is built and maintained by Google employees. While I sleep, Google are making my website faster so that it ranks higher on Google. How do you compete with that? These people are smart. We talk all the time.

We collaborate on stuff. I send them bugs they fix. It’s a really fascinating alien thing. It surprises me that there aren't more SEO people more actively involved in the WordPress world. Because you just get direct access to these people. They're not the people who are designing the algorithms, but they are involved in that ecosystem. There's that.

The other side is stuff close to the core search products. We had little glimpses and little interactions with teams. My world is incredibly frustrating. They are all over the place. They make arbitrary changes to different systems. Like last week without announcing it, they made a bunch of changes to how they consume product data on pages through errors on every page where it was currently implemented.

Our support team at Yoast just gets overwhelmed by people saying, “Yoast has broken my website.” No, Google changed how they consume this arbitrary piece of data. They're really frustrating to work with. We've got some back channels that we can say, hey, help a little bit. But it's all very unofficial. It's a bit weird. Yeah, they are nightmarish to deal with. It just seems like chaos on their side.

Robert Hansen

I have a bit of a love for DeMoss just because it's where I got started. As a result, some people pay attention to the things they say I love, I got introduced to a company called Blekko that was founded by the same guys who founded DeMoss. They're like, “We're bringing it back. We're doing it all over again.”

I sat down and understood a little bit more about what they were trying to do. They had some clever ideas. They we're going forward with it. I was going to ask him, well, how are you backfilling the information? Where's your data coming from? “Well, we have an agreement with Yandex. We share user data with Yandex. They give us search data.” I'm like, “You're doing what?”

Jono Alderson

Sounds perfectly fine.

Robert Hansen

I'm like, “You realize that's very tightly coupled with the Russian government.”

They're like, “Well, it’s no problem. We're just getting search data.”

I'm like, I don't know how to explain this if you don't get it.

Jono Alderson

The deepest darkest needs, secrets, fears, hopes and dreams of people. Fine.

Robert Hansen

I politely told them I could not help them with that project. Later on apparently they got bought by IBM Watson. I have no idea if they still have that integration or not. But I have a similar feeling about all of these incredibly complex systems. From a user's perspective, you're just using this thing over here. You're just using DuckDuckGo. You're just using whatever.

But what you really are using is Google. Because they have an API back to Google. They're handing search data to Google. The IP address, no cookie data, but IP address and search string. Which is everything. I just don't see there being a way that you can really escape that ecosystem unless you fully opt out and go to Bing.

Jono Alderson

Yeah. But even then, I've got an Android phone, a Google TV and a Google Assistant. I've given in. As an end user, the value outweighs the existential fear of all the scary data stuff. Am I being manipulated? What if it gets used? Etc. In the abstract, that still makes me very uncomfortable. But the utility of it all, especially as it all integrates and talks to each other, is so good.

But what's the alternative? Either completely off the grid. I can't use maps. I can't use anything. Can't translate anything. All the stuff. Or hope that Bing somehow at some point innovates disruptively. But I don't see any evidence that they're ever going to invest enough to make that happen. They seem to be quite comfortable to be a distant second place. The features they do roll out are either chasing Google or minor derivatives.

Robert Hansen

I've noticed that. Or even the browser. They've given up. Now it's chrome. They're going to have that same problem where they're always going to have lagging features you’ve just described. I really tried to convince them, don't do this. We were involved in those conversations before they went. I'm like, “Do not do this. You're just going to end up behind them. Way behind, six months, all the time.”

They're like, “Well, it's cheaper.” No. We don’t want to get rid of 100 employees, 1000 employees, whatever it is. One thing about DeMoss and Blekko though. I think people side-eye and laugh that those are human curated systems. I do too. But the thing that I really push back on, is not just that I think it's a good idea or could theoretically be a good idea depending on the implementation, is that Google has that too. They have this enormous content moderation team. It's not they're curating it. But they're curating it in reverse.

Jono Alderson

Validating that the system is returning the things that they would expect it to.

Robert Hansen

Right. If you don't have humans in the loop, this doesn't work at all. You need training data. These systems do not just automatically take care of themselves.

Jono Alderson

I mean, the running joke is that all of this machine learning AI stuff is actually just heuristics and checklists. It's not at all. It's just armies of low paid humans.

Robert Hansen

You and I both know people who have theoretically injected themselves into the search quality teams to change the quality. Or at least to understand how it works.

Jono Alderson

Yeah. I know. Google insists that the search quality rate stuff does not affect the algorithm. But the whole purpose of the system presumably, is to evaluate whether the system is behaving the way that they think it is. If that data says otherwise, of course, that's going to happen. Yes, they're not going to directly feed that back into their algorithm. But it's going to influence what happens next for sure.

Robert Hansen

At a minimum, we know this is at least partially true because we always end up with the most current copy of their rating guidelines.

Jono Alderson

Which is fascinating. They are very dense but they are very interesting.

Robert Hansen

About 100 pages but worth the read. Because that gives you insight into, well, now, I should check these tags. Now I should do this. These are college kids making 10 bucks an hour. But they have a very strict set of things they're trying to look for. If it doesn't fit that exact set of things there, you're going to skirt right through their system.

Jono Alderson

I really love that document because I think it might be the best resource we have in totality that describes what a good web page looks like. Because there's all sorts of stuff like, does this seem to be trustworthy? Is it obvious who wrote it? Is it obvious that the person who wrote it was an expert in their field? Is it full of typos?

There are 100 dimensions of stuff like this that look at or aims to determine if a person searched for toothache. Is this good? Is it right? Is it useful? Is it helpful? Is it trustworthy? It's such a phenomenal resource. If you knew nothing about SEO except that you needed to make a good page, it'd be a damn good resource to use to try and make sure you're getting all the criteria.

Robert Hansen

I agree. That's a good point. Back to your point about the privacy issues related to all these things. I think this is worth enumerating. Because people are like, “Oh, I don't use Google. “ Or they'll say something insane like that. I'm like, well, you use websites. Those websites use Google Analytics. They use Picasa. They use Google ads. Your buddy uses Gmail even if you don't. On and on and on. You really cannot escape it.

I think the biggest indicator that I saw that we have totally lost this privacy war against Google is, they said, “We're no longer going to read your email.” Before that, they had said something like they had 130, 136 I think, signals that they were identifying. They no longer needed the email. They had enough without it. From a signaling perspective for identifying whether somebody would be interested in something or not, maybe it's their search terms. Or maybe it's the 130 other signals that they're tracking.

Jono Alderson

I could get behavioral analysis with abstract labels that we would never align to anything we'd understand.

Robert Hansen

Right. You can say, hey, on my side is completely optimized and done, blah, blah, blah, blah, blah. But it actually turns out it's the fact that they look at these Picasa messages. Or they're using this thing in Google Docs.

Jono Alderson

They say very carefully, whenever they're prompted that, we do not do personalization. They do not personalize the search results or anything around it. But that's a very specific set of wording. They damn well localize it and it damn well changes on your behavior. Yes, they're never using you to... yeah it's just linguistic English at that point, isn't it?

Robert Hansen

There really is no way to disappear from them.

Jono Alderson

No. Certainly not without life changing costs.

Robert Hansen

You’d have to live in the woods. I mean, after listening to this podcast blog, people are just going to stop listen to my podcast and live in the woods. But I don't think you can use any one of those signals anymore. I think you can't say, one, I'm going to block Ads. Therefore Google is not going to see me. You’re using Google Fiber. Or if you're not using Google Fiber, someone else is using some Google Elastic Compute type thing. You're just never going to be able to escape them.

Jono Alderson

Go under the radar. That's hard. I mean, there’s some interesting pushback on that. There's a couple of countries who have now declared that Google Analytics is “illegal”. There's a lot of hyperbole in every action. But maybe it'll start a chain reaction. A lot of those, unfortunately, are quite specifically focused on the handing of data from the EU to the US. Which is one very small part of it. But there's more interesting stuff.

There's a similar thing with Google Fonts. Which typically, if you install Google Fonts on your website, you do by pasting in a few lines of code. It just so happens that you load Google Fonts from a Google domain, which inherently sends them your IP address because you're making a request to a system. That is also now quasi illegal.

The site's being fined, Germany I think, for loading Google Fonts remotely. We're starting to see some recognition of this. But yeah, there isn't a solution. Even if you fix all of these trivial technical details, you still can't escape in totality being part of our system. Definitely not.

Robert Hansen

How would you define the difference between PageRank and RankBrain? What was the big technical delta between those two? I have an answer and an anecdote afterwards. But I'd like to hear yours first.

Jono Alderson

Okay. PageRank was where this all started. Which was the idea that things that reference other things in a network contain some value. You can loosely calculate that. You can work out if one big site links to 10 little sites. Each one of those should get 10% of whatever the value metric is.

That was for many, many years, the heart or at least we believed to be the heart of close to how Google's algorithm worked, independent of all their content analysis and quality metrics. It is very debatable today how significant that is and how much of it still exists. There is definitely a similar concept to that. But it is almost certainly far more sophisticated. To some degree, linking from big sites to small sites is still important. But we have no idea how that works.

RankBrain is a very different thing. It is one part of a whole bunch of named capabilities of Google. They got very keen on this recently. There was MUM recently as well and a whole bunch of others. The best we can tell, this is a post ranking filtering system. This came up in some of the talks earlier very briefly at the conference. They have systems which if I search for dentists, Google just show me 10 links. That's not going to get the best possible result.

The best possible result might be a nearby open dentist, a commerce website selling stuff on demand, a map to something. There's a degree of categorization and filtering and deduplication. It is quite an ambiguous intent. Maybe one of these is going to be information. Or maybe one of them is going to be transactional. RankBrain and various other things related to it, analyze the search phrase to gain as much context as possible based on the combination in the order of the words, and then apply a whole bunch of post ranking filtering.

It will go, we should really have no more than three ecommerce results in this set. Because we're not convinced it's a transactional query. Therefore, elevate a map and a YouTube video. Try and get this best possible blend whilst taking into account the very specific word ordering of the query. All the external stuff.

Then layering in stuff like Google’s MUM things that are multimodal, I don’t remember what it's called, but similar premise that they should do that cross media format as well. They should consume all the images and all the video on the web in all the languages, understand it in some machine learning abstract, and use that to also influence that filtering model.

Robert Hansen

When Bing was first coming out, they had something that was allowing them to train their data. It was like, look at Bing results and Google results side by side. There's this tool they built. You type in whatever. It was a little clunky because they were doing it from their IP address. All the local search stuff was all for Seattle or wherever. Not Seattle. Wherever. They actually have a campus right next to it. I'm spacing on their actual name. I can't believe I can't remember this. Yes. Something like that.

But anyway, the IP address was always the same. It was always a little messy. The results. But it was interesting to watch them. If we've got worse, basically just tells the AI that they have on the back end. Hey, you're a little worse than Google on this one. But it was interesting because there's all this training model. I had this in the back of my head. But anyway, I went to go visit them in Redmond. That was the word I was looking for.

I went and visited them. We’re sitting down and talking about SEO in general. They're like, “Well, what do the SEOs want?” I'm like, “Well, I think if you had to give one thing to them other than better search console stuff, it would be the algorithm. How does it work?” They literally looked at me like I was from Mars. They're like, “What do you mean? What algorithm?”

They didn't have it in their head that such a thing could exist. We just have a black box. It's just a system that decides this is better than this. We don't know how. I can give you the code. It doesn't do whatever you think it is doing. We don't know how it works. It's just this little box. We input data and that data comes out. That's how it gets ranked.

Jono Alderson

I’ve had some of the conversations with some of them. They then have similar stuff to rank. They have the black box. Then a post filtering system that makes sure that whatever comes out is sensible. Otherwise, the thing that ranks the stuff is just completely alien. Nobody knows.

Robert Hansen

No one knows how that works. I think that is more and more. Bing had the advantage of being second, right. They get a lot of the knowledge and a lot of the personnel.

Jono Alderson

They are now solving problems. New problems.

Robert Hansen

They can see how they want it to be solved. They're building a lot of these things. Subsystems to be built entirely off of a black box. Where Google is retrofitting. They're building all this stuff post mortem. Therefore it's taking longer. That would make sense. I mean, that's how things work.

I think that that's where they're going. I think now they're chasing Bing in a weird way. We're just going to have this be an entirely black box where we're not going to know anything about what's happening. Isn't that weird?

Jono Alderson

I’m nervous about the intersect of SEO and machine learning. Especially when it comes to content. People chasing other people. There are so many tools. Again because SEO is hard strategically. It's all about quality, etc. You need to be scanning helpful content. A lot of people look for shortcuts. The shortcut of the day is using machine learning and AI systems to either completely generate or to supplement content writing processes.

The problem with all of these systems, almost all of them and all the SEO tools and platforms, including some of them represented here today. They start from the basis of look at what is already ranking, do some gap analysis versus your site and then fill in the gaps. All these systems and all this content ends up just chasing itself. Because as soon as you do that, the next person does that.

You are recursively optimizing a very small corpus till it's all just word soup. I don't really see a way out of this. If the search engines are doing that too, brands who are just chasing each other to put the right amount of keywords in the right place because Bing Rewards them. This would just create noise. It's really unpleasant.

Robert Hansen

What keeps you up at night?

Jono Alderson

What keeps me up at night? I desperately want all of this to be fixed and better. It frustrates me greatly that it is not. I think the web already has been and can be a phenomenal force for good. But it's still nice, and it's still very early. It's frustrating that it is still tactical. People are lazy or stupid.

Robert Hansen

We already have decades of technical debt already.

Jono Alderson

Oh, yeah. How have we managed that? I'm obsessed by performance. Because there are clearly, there are well lit paths to performance. Yeah. Somehow everything sucks. I'm obsessed about structured data and a connected web. The whole, a slightly less utopian vision than the Tim Berners-Lee Web3 thing. But I think there is an incredible world to unlock with connected data, unstructured data and schema.

I’m deeply fascinated by that. But again, with the way that the ad based commercial web works and the inability of businesses to invest in quality, the thing that keeps me up at night is maybe this is as good as it gets. Or maybe it gets worse. I want to be a force for good. I want to see all this stuff unfold. I want a more incredible connected world of flying cars and whatever. I'm nervous that this is as far as we get. It’s profoundly upsetting.

Robert Hansen

It is. I don't know if it's that desperate yet. But your fears are not unfounded. All right. Well on that lovely note, how do people follow you online? How do people get in contact with you?

Jono Alderson

I'm all over Twitter @jonoalderson. Go to jonoalderson.com. Or you can go find me at Yoast. I'm at various conferences all over the tech scene. You can email me @jonoalderson.com. Or you could probably use Google. Use a search engine. Just Google. Or another Google derived service.

Robert Hansen

Jono, thanks so much for doing this.

Jono Alderson

Thank you. I really appreciate it. Nice.