AlsoAsked Deep Search, AI scammers file fake DMCA notices, Google Webspam Report 2021, Google's new CDN and more

Or get it on:

• Apple iTunes
• Anchor
• Breaker
• Google Podcasts
• Overcast
• Pocket Casts
• RadioPublic
• Spotify
• Stitcher

What's in this episode?

In this episode Mark Williams-Cook and Jack Chambers discuss:

• AlsoAsked Deep Search
• AI scammers file fake DMCA notices
• Google Webspam Report 2021
• Google rolling out desktop support for signed exchanges
• Google's new CDN

Show notes and links

• AlsoAsked's new big update including deep search, improved UI and limitless bulk upload for Pro users
• AI-generated scammers file fake DMCA notices
• Google's Webspam Report 2021
• Google's John Mueller has been replaced by MuellerBot
• SISTRIX's updated Primark analysis article
• Google Search will be rolling out desktop support for signed exchanges soon
• Google launches a new media CDN

Transcript

Jack: Welcome to Episode 16 of Season 2 of the Search with Candour podcast recorded on Thursday the 28th of April 2022. My name is Jack Chambers. And I'm joined by my cohost, Mark Williams-Cook. And today, we'll be talking about AlsoAsked's deep search updates, scammers using AI-generated lawyers to create fake DMCA notices, Google's John Mueller has been replaced by the MuellerBot, Google Search will be rolling out desktop support for signed exchanges soon, Google's Webspam Report 2021 and Google launching a new CDN.

Jack: Search for Candour is supported by SISTRIX, the SEO's toolbox. You can go to sistrix.com/swc if you want to check out some of their fantastic free tools such as their Instagram Hashtag Generator, Google Update Tracker, page speed comparison and tracking of your site's Visibility Index. That's sistrix.com/swc for free SEO tools and sistrix.com/trends if you want to sign up for the monthly TrendWatch newsletter.

Jack: So, Mark, as the founder of AlsoAsked.com, there is a little update for the listeners.

Mark: Yeah. Thank you for adding this because I actually just added this onto our list of show notes about 20 seconds before-

Jack: As I was reading!

Mark: Jack wrote out what we're going to be talking about which is the AlsoAsked update. And I was thinking about this because I always feel weird because obviously it's one of our projects, it feels really self-promotional to mention it and SISTRIX is our official sponsor. And I was like, "So, can I mention it?" But then I was like, "I could just say,'' AlsoAsked don't sponsor this podcast every time."

Jack: It's true.

Mark: And then technically, we're fine. But I did want to mention it because we pushed some new features live on AlsoAsked.com yesterday. And we've had really nice really positive feedback about it, another wave of new users which is always brilliant. So, for those that don't know, we have a tool called AlsoAsked.com which does People Also Ask research. So, you can pop in a keyword or a very specific question which is actually what other tools can't do.

So, things like AnswerThePublic is great for topic research, but when it comes to researching the long-tail what you'd have in your article, that data that they use basically doesn't provide much in the way of results.

Jack: We went into much more detail about six weeks ago when it launched properly. So, if you want to hear a full breakdown of that, you can go back and listen to Season 2 Episode 10 about six weeks ago where we went into a lot more detail when the proper launch of Version 1 happened for AlsoAsked.

Mark: Thank you. Hint taken, I will stick to the updates, which is we've launched a deep search update. And what this means now is there's a redone user interface which should hopefully be a bit more slick. When you do a search, you can now select on a drop-down to do a deep search. And what this will do is it will take you one level deeper in the People Also Ask questions.

So, it's as if you've clicked one level further in. And this actually takes the number of questions you get back and your result from an average of around 20 to over 100. So, it's a huge amount of data for a single-keyword search. On top of that, we also removed the limit which used to be 100 for bulk searches. So, you can now upload a CSV, you can upload 1,000 different queries at a time which means on a single download, you can get 25,000 questions.

So, if you've got your kind of list of topics, keyword research you want to explore, you can work those up in a CSV. You'll get a zip file back which you can then pass out separate CSVs to the content team and it will basically do all of that research for you. So, that's life now. AlsoAsked.com is a freemium model so you can go try a few free searches or it starts at 15 bucks a month.

Jack: So, I know we've been talking a lot about AI stuff recently. We had some special guests on the show last week reading not only the intro but an SEO-themed horoscope for you, listeners. Well, we've got I guess the negative side of people using AI for this kind of thing. And we've had some scammers using AI-generated essentially deep-faking lawyers to trick people into agreeing to DMCA takedowns and notices and paying ransoms and all this kind of stuff.

For those of you who don't know, a DMCA, a copyright infringement notice is something and we've covered this on the podcast previously before as well. It's something that will cause you a lot of legal trouble. It's become a big thing especially in places like YouTube and the music industry and things like that where you're using, basically, licensed stuff. If you're using imagery or music or video or anybody's intellectual property against their agreement, essentially, without their consent and without licensing it for yourself especially for commercial purposes, then you can get into a lot of legal trouble.

Mark: We covered as well before the DMCA tool that is available through Google. So, basically, if someone's scraping or copying your content directly and they ignore you when you ask them to stop, you can actually just directly file a DMCA request with Google who are then obliged to actually take action. And normally, it results in that being removed from their search results quite quickly. And the interesting thing about those requests are that basically happens unless the other person counters it.

Jack: Yeah. It's interesting weird thing we're seeing a lot more of I think and we're going to see a lot more of especially with the power of AI coming around. This was highlighted by Ben Dixon over at thenextweb.com. Of course, links in the show notes, as always search.withcandour.co.uk for all of the shownotes and links. And I'll read a bit from Ben's article here to summarize a few bits for you, dear listeners. Basically, one of his clients was contacted and received a DMCA notice from a questionable law firm called Arthur Davidson Legal Services, sounds legit. And essentially, you have seven days to add image credit to the offending page in quotes with a link to the homepage of her client's website, etcetera. Otherwise, we're required to take legal action.

And as Ben rightfully said, this has been a tactic that has been used before. And you will get dodgy sites to do this. They'd just be like, "Oh, give us a link back to our site." It's kind of a dodgy way of getting links and not best practice, shall we say, for link building to say the least.

Mark: Well, we actually had a couple of years ago, Craig Campbell come to a Search Norwich event to enlighten SMEs and some other SEOs about some of the naughty stuff the SEOs do which ...

Jack: That naughty black hat SEO.

Mark: Yeah, which Craig is very knowledgeable about. And one of the things he said that he had tried with a lot of success was basically just emailing authors when they use image to say, "Hey, that's my image, you need to give me a link." And he covered sometimes, obviously, they reply and they're like, "Well, no, I took this." And he was like, "Oh, yeah, no, it just looks really similar to one I took," and you're doing it as a DMCA is a whole other step.

So, in the UK at least, I don't know about US or certainly outside law but in the UK, it's actually illegal to knowingly file a false DMCA request. So, neither of those things, in my opinion, is a good idea to do, big reputational risks from the first one. And the second one is literally illegal. You're putting yourself at legal risk doing that. So, don't go around trying to claim DMCA rights over stuff unless it's yours, of course, bad idea.

Jack: Yeah. So, this is a law firm claiming to be based in the US. They even had a Boston, Massachusetts-based phone number and seemingly a real building that has real actual law firms working inside of it. So, it seems totally legit. The email ends with references to like Section 512(c) of the DMCA and Professional Signature and looks totally legit. From looking at it, there is a screenshot in the article if you do want to have a check out of that, it looks pretty convincing.

Thankfully, Ben is pretty switched on and was like, "This seems a bit odd." And he went back and contacted them and went back and forth couple of times and got his hackles up a bit and realized something was going on. So, have a look at a couple different ways of investigating this essentially and a few more red flags came up. First looking at the domain record on ICANN, the Internet Corporation for Assigned Names and Numbers, they've been apparently active since 2009. According to the email and according to the website, the domain has been registered earlier in 2022. Red flags straightaway. And it is straight up like arthurdavidson.com. It's a very straightforward domain, seemingly convincing domain as well. It'd be the obvious thing that something a company like that would go for.

Then googling them and looking at new sections and stuff. They claimed to be very high-profile lawyers. They give statistics of 90% success rate, 380 wins from over 400 cases, blah, blah, blah. Basically scare tactics of like, "Look how successful we offer our clients. You don't stand a chance," all that kind of stuff. But nothing came up in the news. And if they had done this, that would be newsworthy. You would be picked up in publications and legal publications and stuff like that.

So, looking at the About Us page on the website, Ben noticed something else as well. None of the people on there are real. All of the photos, despite how convincing they are, are in fact AI-generated.

Mark: That's a red flag.

Jack: That's a definite red flag, yeah. So, using G-A-N or GAN as it's known, so the Generative Adversarial Network, to basically deep-faking people's faces and I know you've played around with this, Mark, for a few different things and similar kinds of things for people's faces, artwork, all this kind of stuff, you get this image training and machine learning stuff that is fascinating and really, really damn convincing to the untrained eye. If you know what kind of artifacts to look for because thankfully there are telltale signs in a lot of ways from this kind of thing.

Mark: For now.

Jack: For now, yeah, until the machines take over and rule the world. But thankfully, Ben was able to notice that and picked up on it and realized, "Yeah, this is complete rubbish from top to bottom, from start to finish." And yeah, very interesting to see how advanced ... I could see this catching a lot of people I know who do not work in our industry who are not tech savvy, relatives, parents, friends who are not in this kind of industry getting caught out on this stuff pretty easily.

And including previous clients I've worked with who are not particularly tech savvy, who are not necessarily legally savvy either, you could very quickly get run into problems with this with such convincing scam. So, keep an eye out, folks, essentially. Make sure you do your research. If you get a hint of a red flag for anything to do with DMCA or any serious kind of legal stuff that people seemed to be just contacting out of the blue, do your research, use the techniques Ben has listed.

He listed some free tools in the article about how he did his research. All of the investigation he did was free just by processing some images and having a look through different things and double checking and searching through certain databases and all that kind of stuff. So, really good resources there from Ben at thenextweb.com, I highly recommend you go and check out the article in full.

Speaking of spam, spam, spam, spam, spam, spam, spam, spam, spam, Google's Webspam Report 2021 is here. The highlight of our year, right, Mark? Yeah.

Mark: I really enjoyed covering the last one. I feel this one ...

Jack: It's pretty dry.

Mark: I feel it's a bit propagandary.

Jack: Yeah.

Mark: I mean, I know there's always going to be that element to Google. But they give some on the surface ...

Jack: Look how great we are kind of stats.

Mark: Yes, very impressive numbers. But then when you think about what they're saying, so for instance they're talking about Spam Brain, Spam Brain launched in 2018. And it's their system to help battle spam AI-based spam prevention. I'll get there in a sec. It's an AI-based spam prevention system. And they've said that they identified nearly six times more spam sites than in 2020.

Jack: That sounds like a lot. Imagine that ... Again, we always talk about this. This is millions and millions. So, we're talking six times millions and millions if not billions of sites.

Mark: I raised my hand to ask was there more spam sites in 2021 than in 2020?

Jack: Or if they just identified more. There's not the clarification there. Yeah.

Mark: Yeah. So, if the system was working at the same efficiency but there was just more spam sites or if there's 10 times more spam sites, they've done a worst job. So, I'm always very wary about these statements out of context.

Jack: The classic SEO thing of ... "The LinkedIn Graph", if you will, just like, "Yep, let's go ...

Mark: And no label on the y-axis.

Jack: Yeah, just the traffic is 100 times higher than it was at that previous point in time. Don't ask about the numbers, no label in the axis, everything's fine. But Google also mentioned there's a major reduction in hacked gibberish spam on hosting platforms which is an incredibly specific metric to pull out but sure, 75% reduction in that as well. And yeah, they're talking about how great they're doing in many ways. And it claimed to keep more than 99% of searches spam free.

Mark: I did some math while you were covering the last topic about that which I was just looking at how many searches Google serve per minute. And I have calculated that if only 1% of their searches have spam in, that's approximately 54,720,000 searches per day returned spam.

Jack: That's a lot. Again, when we're dealing with these astronomical numbers, that is a lot.

Mark: So, the reason I'm kind of a little bit lip-girly about this is because like we talked about in the last episode and the experiment I'm running at the moment, I've seen more spam in Google over the last 12 months than I have done in many, many years.

Jack: Especially when we're talking about all this AI-generated stuff. If you're counting that as spam and I would to be honest if it's not serving ...

Mark: For sure.

Jack: If you're not serving the user intent and you're just churning out crap like the experiment you've been running and I've seen quite a few other people tweeting about like, "Oh, yeah, this AI-generated site has now hit half a million visitors a month or whatever." And it's just they have no quality control for any of this kind of stuff. And is there a way of picking up on this kind of stuff that Google has? They claim they are picking up on things like scams as well, the fact that we just touched on scams there as well, resulting in a 40% reduction in scammy results, whatever that means. So, yeah, big numbers claimed by Google but I think you're right, Mark, because AI is becoming so sophisticated and clever people are using it in clever ways now, it's difficult to try and keep up with that I think.

Mark: Like Google have said, it's a bit of cat-and-mouse game.

Jack: Yeah, of course.

Mark: And certainly, the mouse has got further away from the cat. These spam sites because I've been looking into this a bit more because I'm interested in what's working because I want to understand why it's working. So, I need to see what's working on, that kind of stuff and why it's working, then apply that actually just within Google's guidelines.

And the interesting thing I've noticed is that some of ... Well, a lot of the spam sites I've seen and by spam, I mean the ones that are just pumping out pure AI-generated content which is hit and miss, some of it's like, "Wow, that's a great answer," and other thing is like, "This has got the wrong end of the stick."

Jack: Yeah.

Mark: They're not even on any particular topic. So, I've seen these sites with the quarter, half-a-million, a million, two million pages and they're just answering random questions like, "How long does it take to cook this egg thing?"

Jack: And then the next one will be a video game. And then one will be like, "How to brush a dog?" And be like, "How are these things?" How is Google or any kind of crawler trying to understand what this site is trying to be authoritative about? And we talked about this, of course, recently as well about EAT and authoritativeness and trustworthiness, all that kind of stuff bringing expertise to something.

If you're just covering random nonsense over and over and over and over again, how can that get through the filters? How can you get away with that stuff? There clearly is some stuff that is getting through Google's filters because we've got anecdotal evidence of it on Twitter and things like that. And through the experiments you've been doing recently, Mark, literal evidence from what we've seen watching your pages get indexed on your experimental site looking through Bing churning through stuff and being like, "Bing's happy to index." And then looking at how Google is slightly slower, but still you're getting indexed quicker than some total legit sites I have worked on in the past and I'm currently working on. It's bizarre.

Mark: Yeah. It really flies in the face of a lot of people talk about topical authority because it's just random topics. And I know that Google will deal with many sites like Reddit, like Quora are the two main examples where they're user-generated content and they just cover all topics and new sites as well. But this is just really low-quality sites with no, as you say, no expertise, no history, just churning out maybe the specific stuff.

Jack: Based on the opposite of what EAT stands for or what we've been led to believe is so important by Google.

Mark: I think it's great that they're focusing maybe on stuff like hacked links because again, that's crossing the line in terms of, obviously, that's literally again illegal.

Jack: Yeah, spamming and scamming. You've got two different extremes there.

Mark: Yeah. So, I think Google's doing the right thing if they're focusing efforts there because that's protecting users, that's protecting webmasters, that's protecting businesses and stopping literal crime. Whereas it's not the end of the world if you get a spammy result that isn't quite right and there's just some ads on it. And you go, "Uh," and you move on. But yeah, definitely cat and mouse, definitely is highlighting some I think gaps that have been created in the algorithm for this type of content. Yeah, we'll put a link to the spam report so you can have a read through that.

Jack: Here we are at the midpoint in the show. A little update from SISTRIX and we've got a specific update. We touched on the Primark new informational website a little while ago and we're back on the Primark website. Are you excited to discuss Primark again, Mark?

Mark: I love chatting about Primark's SEO.

Jack: We were genuinely interested. And it was kind of a thing I'd not really, like I tell at the time, I had not really considered how important and how different this could be for trying out new things in a weird non-eCommerce but an eCommerce space kind of thing that we talked about.

Mark: So, just to catch people up if they missed the episode where we talked about Primark, basically they built a new site. It looks like an eCommerce site, but of course being Primark, you can't buy from it. They're doing a migration. And we're interested to see will Google rank now product pages even though you can't buy from the site. So, there's that mismatch of user intent for unbranded queries, "I'm searching for fluffy socks, will Google rank it even though I can't buy them?" So, that's why we find this particular case interesting.

Jack: Yeah. And it is pretty much an eCommerce site minus the Add to Cart button. I think was pretty much a quote from you last time we talked about it, Mark. You get prices, you can get stock notifications, you can get all kinds of stuff, everything except actually purchasing things without having to walk into the store yourself. And we've actually got an update from Steve Paine from SISTRIX who wrote the original article. The article has now been updated. So, if you do click on that link in the shownotes, this is the new updated version of that. And I've got a quote here from Steve specifically about what we're seeing and what he thinks is going on, essentially, from the data we're seeing from SISTRIX and how Primark are trying to manage everything in a careful PR kind of way.

So, here's the quote from Steve. "It's currently a net loss. But the company, the parent company that is, is saying Primark have seen a good early reaction to its new UK non-transactional customer website with traffic doubling in its first two weeks according to its parent company, Associated British Foods." Back to Steve, "This is a classic measurement by traffic kind of problem. You can do some PR. Everybody's covering like, 'Oh, wow, Primark has got a new site. They're a big brand. People are going to cover this on big news sites. Traffic is going to go out from that naturally.' Is that actually sustainable though with a net loss and visibility?" Steve reckons, "No."

Jack: And I would tend to agree with Steve. I think this is a very interesting case of a really big brand here in the UK doing something different. And as you said, Mark, not matching user intent which I think could be a cardinal sin for there and kind of the nail in the coffin for a lot of their visibility and a lot of their traffic.

Mark: So, the visibility they did have that we could see from SISTRIX prior to the migration was pretty much all branded so ...

Jack: It was heavily, heavily weighted as branded traffic, absolutely.

Mark: And it still looks like with the latest data, we can see pretty much all the things they're ranking for have the word Primark in everywhere. So, it would make sense that search traffic would go up as well from that digital PR point of view because people are just searching more for Primark branded terms. There's still no movement really on the unbranded stuff. If anything, it looks like, as Steve said, a net loss. So, yeah, I think they're definitely trying to polish that result.

Jack: That's a very diplomatic way of putting it, Mark, yeah. And another update from SISTRIX as well, we have new features for keyword gap analysis. We've got two new features coming through on the gap analysis side of things. We have a feature called Compare Keywords where you can see at a glance how well your domain is performing at a keyword level in comparison to your competition.

And this feature is now revised and now suggests potential competitors as you type. So, kind of auto-completing and suggesting, "Oh, maybe you hadn't thought of that competitor," or, "There you go. Here's an example of people who are likely to be crossing over with you in keywords and competing with you on a search-by-search basis. If you rank better than your competition, the rankings are highlighted in green. If your competitors rank higher than you, they're highlighted in red." So, you can get a quick at-a-glance kind of idea of how you stack up against your competition.

You can then also use a filter called Common Keywords that lets you narrow down the keyword results to see minimum number of competitors that are ranking, the most number of competitors that are ranking, all that kind of stuff. And as always with most of the SISTRIX kind of stuff, you can filter extensively and really get down to the nitty-gritty of the data there as well.

The other updated feature is Unused Keywords which then shows you the keywords for which competitors already rank for but which you do not. So, looking at where the gaps are for you in your content.

Mark: That's one of my favorite features. It reminds me of tools like Majestic, you have Clique Hunter which basically does the same with links so you can put yourself, your competitors in and it will show you where all of your competitors have links but you don't. I think it's really important for SISTRIX to have this from a content point of view because it really highlights where if three of your competitors are ranking for this key phrase, it probably means they're covering some topical things that you're not that you need to be looking at and prioritize because that can actually impact your rankings as a whole. It's a really, really, really cool feature there.

Jack: Exactly. And again, with the common keywords tool, you're totally right, Mark, you can then filter it down even more and narrow it down to where your competitors rank in the top 10 and top 20 and get even more granular there. And see, "Yeah, I should probably be covering this kind of stuff because a lot of the competitors are covering it as well." And if you know you can do better than them and create better content, then you can compete with them on those keywords and beat them at their own game in terms of content. And see essentially where gaps in your content are.

So, some fantastic updates and some really great content ideas hopefully coming up soon from you guys using SISTRIX. And you can find all that information of course by going to our shownotes at search.withcandour.co.uk. There'll be a link for the Primark article and the updates for the features there as well. And sticking finally with the last piece of AI-generated content, I'm not saying the content is generated by AI, but in this case it is because Danny Richman is back, ladies and gentlemen. We talked about him a few weeks ago with this fantastic English to regex tool. And Danny is very clever and doing some very clever stuff but also likes a good laugh, thankfully.

Jack: So, we don't need John Mueller, the search liaison from Google anymore. We have MuellerBot who can seemingly just answer any question you need him to because ... Godspeed to John Mueller. He answers a lot of questions on Twitter that he's answered a million times before. And I have seen to the point now where Danny is replying to people asking John Mueller questions with gifts of his misgenerated answers from his artificial John Mueller, and they're really damn good. They're very convincing and it's hilarious.

So, yeah, if you don't know Danny Richman like I said, we talked about him a few weeks ago. He created some amazing GPT-3-powered scripts. A lot of them are available for free for you to test out and experiment with and build upon yourself on his website. Of course, links in the show notes as always at search.withcandour.co.uk. But yeah, I think Danny started this off as a joke and he says it in his thread of tweets here, "Originally intended as a light-hearted joke until it became apparent that MuellerBot's responses were, in some cases, unnervingly accurate." And they even have and I don't know how this works, John Mueller's wry Swiss wit there somehow. It's truly brilliant and something I will never get tired of seeing Danny replying to people with silly SEO questions on Twitter.

Mark: So, I think this started because there was a thread where Paul Shapiro said to John Mueller, "Could you get a bot trained to answer these results?" And John said, "I would pay for that." And I think Danny ...

Jack: So, Danny did.

Mark: ... saw that and went off and made it. And I talked to him a little bit about how it gets these answers. So, the actual answers it's getting ...

Jack: Is from John's tweets, isn't it?

Mark: They're not.

Jack: Are they not?

Mark: They are the general understanding of open AI. And he did tweak the model on only a small sample of John's tweets to change how they were phrased.

Jack: The tone of voice.

Mark: Yes. But the core knowledge which is asking really specific things like, "Can you optimize for RankBrain?" Or here's one, "Is domain authority a factor in Google's ranking algorithms?" And it's generating answers saying, "In general, Google doesn't evaluate a site's authority so it's not something where we give you a score on authority and say this is a general score for authority on your website. That's not something we will be applying here." That's pretty much correct.

Jack: So good, yeah. Almost every single one I've seen has been more or less the correct answer I feel. And not to speak a word of John Mueller, not at all, he's fantastic but this is game-changing-putting-me-out-of-business kind of stuff, right?

Mark: Yeah. Will canonical tags help with the indexing of my site? Maybe. So, I think if you're having indexing issues where some of your content is not being indexed for some reason, then using a real canonical can be a way to help us understand that this is the content that you want us to index. So maybe, that's something worth trying if you're seeing these sorts of indexing.

So, it's really interesting that you could look at ways of actually implementing this and then just having a human check over answers. So, an Ask Me Anything About SEO run by GPT-3 on pooled knowledge ... The problem is obviously that it's going to give some cracking real answers at some point that will slip through the cracks of the price of having a false positive if you lie answer would be high.

Jack: Especially if you're giving it the authority of like, essentially, speaks on behalf of Google, on behalf of John Mueller.

Mark: But there are companies. So, there's one actually in a building that for years was building a system. And the idea of that system was to essentially take all of a company's knowledge from all the individual people that work there and put it into a system that you could just ask conversationally questions to. And they've been working that technology for a long time that, basically, we've leapfrog there with GPT-3 and open AI. So, you can give GPT-3 a specific corpus of documents to be used as its knowledge base. So, if you have your own manuals, documentation, wiki intranet, you could do it on that. And that would maybe stop it going off piece.

Jack: Yeah, incredible stuff by Danny as always.

Mark: Yeah, PR wise for Danny Richman and what he's doing with GPT-3, it's really cool. I think John enjoyed it as well.

Jack: Yeah, two of my favorite people to follow on SEO Twitter. So, we'll link to the Twitter thread and Danny's website there as well so you can go and check all that in the shownotes.

Mark: Okay. Google is going to be rolling out desktop support for signed exchanges soon. This is a topic we've covered over the last couple of years actually. So back in April 2019 in Episode 6, we first talked about Google was going to start showing search results for AMP with the hosting domain rather than Google's properties. And then fast-forward just over two years into the future in September 2021, Cloudflare released which is still in beta that kind of one-click signed exchanges. So, this meant if you are hosting your site on Cloudflare, they can basically allow Google to directly cache your website for mobile and it would be super, super, super quick AMP kind of style. That's something we've been pushing a lot of our clients on Cloudflare to expand it with because the difference in performance is huge.

So, the update that we've got for you today is Devin Mullins who is a software engineer at Google working funnily enough on signed exchanges. As I said, Google Search will be rolling out desktop support for signed exchanges soon. For almost all sites, no action is needed.

Jack: Hurray. Lovely. That's what we like to hear.

Mark: And that's super cool.

Jack: Thanks, Devin.

Mark: So, they're just going to do it. Sites serving different HTML based on the user agent header will need to opt out by adding a meta tag to your page. What does that mean? And he's also said that, basically, sites using responsive web design or separate mobile desktop URLs don't need to take action. So, to just dive a little bit into what he's saying about dynamic serving, so this is if you are showing different results to Googlebot for instance, to users, so you are checking the user agent so whoever is requesting the page, the URL, who they are. And so, this is quite common, we've talked about it before. When for instance you might have a lot of client-side JavaScript on your site that bots would struggle with, so you might have a slightly different dynamically-rendered version or server-side-generated version or a static version that you serve to bots which is what Google recommend in certain instances.

So, here it said, "Science using dynamic serving by varying user-agents will need to annotate their pages as mobile or desktop only." So, there's some documentation there that we'll link to in the show notes, search.withcandour.co.uk. And basically, it's just a meta tag you can put on your page that specifies whether it's mobile or desktop.

The thing they're trying to avoid here is otherwise desktop users might see the mobile version of the page so they'll end up basically caching the wrong version.

Jack: I have experienced that myself and trying to navigate a desktop website on your phone is one of the most frustrating experiences of using a smartphone I think I've ever had.

Mark: Yeah, this is the other way around. And actually, I encountered this yesterday on a site that basically it has a separate mobile and desktop version. And because the indexing is switched to mobile-first, all of their mobile pages are cached for desktop results.

Jack: So, suddenly, there were burger menus everywhere and then long scrolling.

Mark: Yeah. And it's kind of like the mobile version of their site is very like, "Oh, we've just kind of crowbarred this and so it roughly works on a mobile." So on desktop, it's just awful.

Jack: Interesting.

Mark: Yeah. So, this is what they're trying to avoid. Google said, "We're reaching out individually to these sites," that they found to be using dynamic serving but they've put this announcement out because I imagine there will be people they missed because the web is big. But, yeah, exciting news because it means actually that lots of people should be getting a kind of free performance boost.

Jack: Yeah.

Mark: That's the upside. And I guess there are some monopoly-type downsides. We'll talk about them later at another point. Let's leave it positive: free performance for everyone.

Jack: So, you don't need to do anything for it. You're just going to get a boost in performance from Google.

Mark: There we go.

Jack: There you go. Lovely.

Mark: Cool. I wanted to finish off this show just talking about again a new thing Google has launched. And in this case, it is Media CDN so Media Content Delivery Network. Google is getting into the content delivery network game. It's I guess not fully related to SEO or PPC. It's kind of related to PPC. But I think it's interesting so I just want to cover it briefly because I've got a feeling it's one of those things that will probably get quite big in the future. I saw this through Dr Marie Haynes on her Twitter, brilliant person to follow, which was-

Jack: A regular mention here on the show.

Mark: Yeah. And another one of the people that we end up mentioning at least once every few months. And it was a TechCrunch article entitled Google Launches Media CDN to Compete on Content Delivery. So, this was announced at the 2022 NAB Show Streaming Summit which was on the 25th-26th of April in Las Vegas. There is an accompanying Google Cloud Blog post which again, we will link to in the show notes. I'm just going to give you basically the super short version of this article which is that Google have got this CDN. And of course, being Google, the infrastructure behind it is huge. So, they've said it's got reaching over 200 countries and more than 1,300 cities.

Jack: I couldn't even name 1,300 cities.

Mark: Yeah, right? It includes all the kind of latest technology out of the box so quick which is HTTP/3. And Google says, "With multiple tiers of caching, we minimize calls to origin even for infrequently access content." And that's Google's VP Shailesh Shukla and he wrote that in the blog post. "This alleviates performance or capacity stress in the content origin and saves costs." Basically, is Google saying, "Hey, we're really, really, really, really good at performance and caching optimization because we own a couple of properties like you might have heard of it, YouTube." And one of the stats actually in the Cloud Blog Post which I hadn't seen before was so streaming video accounted last year for 53.7% of all internet bandwidth traffic.

Jack: That's a hell of a number, more than half of all internet bandwidth. I am definitely a contributor to that. I'm getting to the point where I'm listening to some podcasts on YouTube now and things like that. When I'm at home, when I'm cooking, when I'm out shopping, I will often have YouTube things playing in the background. So, I've seen my YouTube use statistics and they're terrifying sometimes. So, yeah, I'm not surprised. And especially with things like lockdowns and things happening recently, unsurprising that more people are working from home. You have that stuff playing in the background. So, yeah, unsurprising but still astonishing that more than half, 53.7% is of all the total internet bandwidth, it's absolutely crazy.

Mark: So, TechCrunch had a quote from Eric Schmitt who is a Senior Research Director at Gartner. Do not be confused with Eric Schmidt which was the former CEO of Google. That's not who this is.

Jack: So, the CEO of Google is Schmidt with the D-T at the end. This is Eric Schmitt with two Ts at the end just to clarify that if you're searching for these people.

Mark: Yeah. So, the quote is, "Media CDN marks a further expansion of the Alphabet empire," Alphabet obviously being Google's parent company, "in this instance by commercializing the pipes that YouTube uses to deliver streaming video." I thought that was quite interesting because one of the technical achievements that Google have had to make is the actual infrastructure behind YouTube which must be horrifically scary because out of all that video, they probably account for a big chunk of that.

Jack: Yes, a massive amount I would assume.

Mark: So, he told TechCrunch via email, "TV and video content providers that choose to build on Google Media CDN technology can expect top-notch scalability. On the flip side, prospective customers must weigh the risks of becoming technically dependent on Google for advertising, delivery, and ultimately perhaps even commercially dependent on it for ad sales and measurement, as Google builds out linkages between its cloud offerings and its advertising software products."

Jack: Not to spinoff we missed too much, but this is definitely something we touched on with the, how do I word this, the alternative to GA4 we were talking about and how much we all rely on Google and Alphabet products in general, whether you're working in the SEO industry, chances are you are if you're listening to this podcast and we are as hosts of this podcast.

But like I said, my partner, my family, my friends are all using YouTube. Google has become so synonymous with it is the verb to search something now. You don't go and ‘search’ something, you go and ‘Google it’. It's become this thing. And yeah, I think there is a definite ... I'm not saying like, "Oh, there's a case for monopolization or anything like that," but it is pretty scary that you get these entire industries and entire companies built around being completely dependent on technology built and owned by Google, advertising, revenue streams completely built and owned by Google and all that kind of stuff as well. And we're totally guilty of that working in SEO and PPC. That is a key, key part of our industry whether we like it or not.

Mark: I mean, this isn't my expert area but I see this as a play for the Google Ads has been pushing hard now for the last couple of years for YouTube video advertising. And if they've got this Media CDN that is geared up really well to do video, they can get people streaming live events using their infrastructure. They're going to have that technology to inject ads to have user data around that.

So, it essentially gives them a large again third-party video advertising programmatic network. Because if you've got these new or if you have even ... When I say new, these older companies that are moving on to digital say someone like Disney for instance, that why would they bother trying to build their own infrastructure if there's something that is programmatic, it's got all those tools, measurement, everything in it? As Eric says, yeah, it's just that thing about weighing up your kind of vendor-locking reliance on one partner.

Jack: Yeah. And talking about how Google wants to add advertising on everything, I just got up a quick little statistic here that $209.49 billion was Google's ad revenue in 2021. Nearly $210 billion in advertising revenue from YouTube ads, Google ads, display ads, product listing, service offerings across their entire ad network basically. It's big money. And it's a big key part of Google's revenue as a company as well. So, the more places they can get their ads in, the better it is for them.

And I know this is something we've been talking about with our PPC team here as well. There's been some serious pushes to automate a lot of stuff and make you spend more so then they make more money from you spending more in Google ads and stuff like that as well. So, yeah, go in with a bit of caution.

Mark: Just looking over your shoulder there, I was always kind of quoting people that are still around the $120-130-billion-dollar mark.

Jack: Three or four of year's update.

Mark: Yeah, that jumped between 2020 and 2021. That's like going from $150 ...

Jack: Less than $150 to over $200 billion.

Mark: That's like 30% growth year on year on ad revenue which is not many people are binging it. What's happening to other people binging it?

Jack: It's more than the 30% growth. Yeah, exactly. I wonder if that is because of a lot of the automation and performance stuff we've seen pushed in Google ads. That's going to contribute a little bit obviously. But even things like I was saying, more people using YouTube, more people using these platforms. Everybody is working from home and more connected than ever and all the kinds of typical stuff you hear from every tech resource in the world.

I know I'm treading old ground here at this point, but it's an interesting thing to see just how much money. And as you said, Mark, that there's a very clear kind of graph and steady growth there, but there is a pretty serious jump up from 2020 to 2021. And maybe we'll see that again there in 2022 when we have the statistics in a year or so time in February 2023 or then get the full-year statistics then we'll see maybe we'll cross the $300 billion threshold or we'll go crazy.

Mark: I knew there was a reason why my friend who works in PPC literally has a t-shirt that says, "Wasted spend is Google's friend."

Jack: Yeah, that doesn't surprise me. And like I said, we talked about it with the PPC team here at Candour quite a lot. We're aware of this and I'm sure plenty of you are, listeners, as well. But yeah, keep an eye out on what Google is spending and the optimization recommendations they're giving inside the ads dashboard and all that kind of stuff. A lot of it is just them trying to get you to spend more money so they make more money. So, keep an eye out for that kind of stuff because Google's a corporation at the end of the day and Alphabet is one of the biggest companies in the history of the world. So, don't just go trusting them with a blind eye like, "Yeah, they're optimizing stuff." Like we said, they give us stuff for free. They give us a performance boost on our website. What can possibly go wrong? They're still trying to make money. They're still one of the biggest corporations in the world and probably will continue to be for quite a while. So, take everything with a pinch of salt and a hint of scepticism I would say.

**Jack:**** So, that's all we got time for this week. We will be back next week. Monday, the 9th of May, we will have our very first in-person guest of Season 2. The aforementioned, he's been mentioned a few times on this season. Steve Paine from SISTRIX will be here, ladies, and gentlemen, next week. We will talk to Steve about his career so far and get some real in-depth insider knowledge on SISTRIX how they work and how they came to work with us and you'll understand a bit more about our sponsor for this season. So, look forward to that next week. Until then, have a lovely week. And thanks so much for listening.