Candour

Episode 92: 2020 Christmas special with Judith Lewis, Chris Green & Kristina Azarenko

Play this episode:

Or get it on:

What's in this episode?

In this episode, Mark Williams-Cook is joined by:

They'll be discussing what we've seen in SEO during 2020 and what they think will be important for 2021 onward, including Core Web Vitals, AI, and good old-fashioned fundamentals!

Transcript

MC: Welcome to episode 92 of the Search with Candour podcast, recorded on Thursday, the 17th of December 2020. My name is Mark Williams-Cook, and it's that time of year. It's almost Christmas, so it's time for the Christmas special, which means we're getting a few other SEOs in to talk about what's been going on in 2020 in SEO and what we think might happen in 2021. Today, we're going to be joined by Judith Lewis, Chris Green, and once again coming back, Kristina Azarenko. We're going to talk through everything, all the pains and woes of 2020, from indexing issues to what might happen with Core Web Vitals and GPT-3 in 2021.

Before we get going, as always, just want to tell you this episode is very kindly sponsored by Sitebulb, which is a desktop based SEO auditing tool for Windows and Mac. You know, hopefully by now, if you've listened to previous episodes, I've been a user of Sitebulb for a long time, so it's an absolute pleasure to talk about them. I normally talk about different features they offer each week, but I actually saw something this week on Twitter, which I just want to bring up and really reflect on the impact Sitebulb is now having within the SEO community.

I saw a Tweet on the 16th of December, so actually yesterday, that said, "Someone's looking for some advice on a WordPress site. Not sure whether this is a good or bad idea," and it was about robots.txt blocking some of the WordPress content folders, the plugin folders, and the upload folders. They were wondering what the impact of this might be, and someone actually just pointed out the best place to look for this is actually in the feedback that Sitebulb gives you when you do an audit, and it uncovers these kinds of URLs that are blocked.

Sitebulb literally has some warnings for things like disallowed style sheets, disallowed javascript files, and it explains very clearly things like CSS files that are disallowed or javascript files that are disallowed may affect how search engines render content. Then obviously, if they're unable to render your pages, users see them and they may not understand exactly what the page is about and you might not actually rank as well as you hoped. It just really struck home to me that that's the best advert, apart from this of course, that Sitebulb could actually have, where you've got the SEO community just referring back to their tool as a source of truth.

Good news is there is an offer, if you listen to this podcast, for Sitebulb. If you go to Sitebulb.com /swc, you will get an extended whopping 60 day trial. There's no credit card or anything, so it's totally free. Download it, give it a go. If you like it, it's there. I strongly recommend it.

Welcome to this episode, and we are very, very lucky to be joined by Judith Lewis, Kristina Azarenko and Chris Green. Hello, everyone.

CG: Hello.

JL: Hello.

KA: Hey, hey.

MC: This is actually the first time I've been able to see my guests, as we've decided to join up on Zoom at Judith's suggestion, which I'm loving as well. You won't be able to see it, but Chris is sporting a rather nice Christmas hat, I guess nice.

JL: Very festive.

CG: All right, I'll take that as a compliment.

KA: Yeah, it's amazing. You all should be jealous because you can't see it, but it's amazing.

MC: I'll tell you what, Chris. I'm going to take a screenshot and we'll put it on.

CG: I also have a robotic T-Rex jumper as well. I don't know if you can see.

KA: Wow. Oh my God.

KG: It's entirely for me, not for anyone else.

MC: I'm going to let you guys, one by one, introduce yourselves. Just tell us a little bit about yourself, your background in SEO, as there might be some people listening to the podcast that haven't heard of some of you. Judith, if we start with you please.

JL: Sure. I'm amazed at the Judith/Julia Logan. Irish Wonder and I may both be black hats, but we are different human beings. I know we're rarely seen at the same place at the same time, so it could be called into question, but yes, we are separate human beings. I am Judith Lewis. I'm founder at DeCabbit Consultancy. Why is it a consultancy? Because I've been doing this for almost 25 years, and if I didn't love it, I'd have left the industry by now, like many of my illustrious cohort back in the early 2000s, late '90s.

I've done in-house, I do consultancy, I've done agency. I've done, before I was married, many people too. Yes, I am that person. I did work in the porn industry, so porn. I've done casino, I've done pharmaceuticals, so I've been in every really racy industry of the SEO world. I've also grown up through the PPC era. I've been in SEO for more than almost a generation, so almost 25 years, which means that I have been alive longer than a lot of people, but not only that, I've been online since '85, so I've been online longer than some people have been alive, and I've been in SEO longer than some people who are doing it have been alive. That's my claim to fame, but I do love it and I continue to do it. I'm now a consultant. I come in, tell you what to do, and then *** off.

CG: Living the dream.

MC: Wow, good luck following that intro. Judith, just to explain the reason in my notes, why I have Judith/Julia Logan, was just to remind me to tell everyone that the last time I actually messaged you directly on Twitter was to give you a compliment on the videos you had done with Julia/Irish Wonder because, in terms of SEO content, I think I said to you at the time, I've got to the stage where I'm quite picky now with what I spend my time looking at and reading. I really enjoyed that series you were doing with it. How can people find that, if they want to find that on YouTube?

JL: Wow, so on YouTube it's called the SERP Show and we do keep trying to record more episodes of the SERP Show. I think we did Viagra, and we've done CBD oil, and things like that. Oh, I think we did a Bingo one as well, because there were some interesting things that were going on with a particular Bingo, let's say supplier both online and offline of Bingo joy. But yes, we have more in the pipeline. We just have to find the time to do it. It's easy to do this because you've done all the prep work, but our SERP Show, it's generally about four hours of prep work.

MC: Wow.

JL: Then we do about an hour to an hour and a half of recording, and then there's post production, so it's quite a palaver, which is why there are so few episodes.

CG: Julia, I could have sworn you were going to say we did Viagra, we did CBD, and then we recorded the podcast. I thought that was what you were going on about.

JL: Not as much Really. It's just CBD oil, it will relax you, and Viagra is a vasodilator, so we'd just be extra relaxed.

MC: Wow. Kristina, welcome back. You've been on the podcast before, you've spoken about E-commerce. It wasn't quite like this. It's gone off the rails since then.

KA: Honestly, I have nothing to say about porn or Viagra, but I've been in the industry for 10 years. It's so funny. When Judith was talking about how long ago she has been in the industry, I was like, "Wow, I got my computer when I was 15 and I got the internet when I was 18," so I was just a quick learner, but obviously not from the very first taps of my life.

Yeah, so I've also worked on different angles of SEO, so to say in the agency environment, in-house, and now I'm a consultant for a year and a half. Honestly, I also could have just forgotten about that and returned to any 9:00 to 5:00 job, but I'm actually loving it. It's a very interesting path of self discovery. Also, when you choose to work, you really choose the clients to work with and you really are so invested in the results. I just love it, so I chose to work mostly in the eCommerce industry helping online stores, but I also do a lot more dealership websites, which is really interesting. I'm also trying to get as much information out to share with the community and I love doing this as well.

MC: I think recently, and I was going to ask you about this, you've just published a new or updated book on SEO for eCommerce. Is that right?

KA: Yeah, exactly.

MC: And you've just released a Chrome extension.

KA: Yes.

MC: Do you want to just tell us very quickly about those and where people can find them?

KA: Yeah, sure. My eCommerce SEO ebook was called eCommerce SEO Mastery. It started two years ago as a checklist, but then it grew into really big and really helpful ebook for everyone who wants to get into eCommerce SEO. Yeah, people find it really helpful, so you can find it on Google eCommerce SEO Mastery, or you can go to my MarketingSyrup.com website and there's an ebook link in the navigation. You can click there as well.

KA: Yeah, and the second thing is Chrome extension. It's a completely free extension. You don't need to even leave your email address or anything like that, but it gives you really quick insights of on-page SEO in just one click. You can see schema, you can see status codes, or link chains, resurrect chains, title tags, and everything that you need. It's called ... Yes, I just forgot to tell this. It's called SEO Pro Extension. You can just Google it and it will be the first result.

MC: Don't worry. To everyone listening as well, you'll find links to everything we're talking about at the show notes, which are at search.withcandour.co.uk. Last but not least, Mr. Chris.

CG: Hello. How's it going?

MC: Hello. Yeah, beat those intros. Beat those intros.

JL: It's a competition now.

CG: Oh man, mine's quite pedestrian by comparison. I mean, I don't know how you characterize this. I think I've been paid for messing around on the internet for well over 10 years now, but I've been messing around on the internet for as long as I've had access to a computer, so a while. Not quite as long as Judith. You kind of beat me on that side, but for a long time.

MC: What did you say, '82, Judith, is when you were on the internet?

CG: '85, I think she said.

MC: '85?

JL: '85. Well, '85 because I was on an internet, so I was on a network set of computers through a Gandalf server connector with dome terminals, so our network went out to a separate database. Then I dialed up into BBSs and things like that.

MC: I think there's a whole other interesting story there. Sorry, Chris.

CG: That's more interesting. I mean, yeah, your internet story predates me by a few years anyway. What do I do? My job title is Head of Marketing Innovation at a small agency in Essex in London called Footprint Digital, but effectively that means I'm getting paid to play around on the internet still. I think SEO is my core, but actually I spend a lot of time with analytics, paid search, CRM, email, managing digital teams. Most of what I've done is in agency, but I've done a small stint in- house and realized that I prefer agency vastly because ... I don't know, because I'm a glutton for punishment I guess. I'll put it in that sense. It's kind of a punishment. I mean, this is December and I'm in an agency, so it's one of those spaces.

CG: Yeah, I just love not knowing the answer to a question, but the journey of finding out what the answer is. I think out of all the areas and jobs that I've worked in, SEO is the one that's been the most littered with questions. You've also got to learn to be slightly okay with unsatisfactory answers as well, or at least put up with them because, if you must know the answer to any one given question, go and do something else. That's fundamentally the thing, but if you're willing to try something new and work around a problem, then that's kind of the core stuff. I play and get paid for it is how I see it, which I think is a win/win.

MC: That's a good way to look at it. You've spoken about search and Orange as well before, and site migrations. I know you've helped us out with CISTRIX training and you're my skipping nemesis. Chris, I don't know if you've seen, is absolutely incredible at skipping, so it's something I started during lockdown as well. I was like, "Hey, I'm getting good at this," and then I saw a video of Chris doing these double under things, like jumping, like out of Rocky collage or something.

CG: I'm like a budget Rocky. I've never knowingly been a nemesis of anyone, so I don't know how to take that.

MC: In a good way. You're my motivation.

CG: If you rotate your camera to the left, is there like a dartboard of my face on it? Has it got to that level yet?

MC: No, no, it's downstairs.

KA: It's downstairs.

CG: It's in the garage, yeah.

MC: This Christmas episode, like last year, we have gathered some very smart people to talk about what's been happening this year in SEO and in search, and what we think might be happening next year so that the following year we can look back and then see if we've made an idiot of ourselves or not. Last year we had a collection of people, and I actually checked the notes just before we started here. I don't think we were too far off. We talked about voice search and we actually talked about how we didn't think it was going to be much of the bigger thing this year, and I think that's a fairly okay thing to say. People are talking about because there's not a particularly easy way for people to make a lot of revenue from it right now.

We talked about better understanding of intent with BERT, increased automation, especially with PPC, with Google, and we talked about Google may be in the next couple of years moving away from actually being keyword led with PPC and just trusting in the big box as Google. Certainly we saw that with the sort of surprise removal of the search terms data from Google Ads, which I don't even want to talk about that today because it just upsets me so much. We talked about more machine learning/AI in search.

There's a quote actually from last year that I'd written down, that basically Chris just said as part and parcel of being an SEO, which is moving more to the situation of doing something for SEO and then saying, "Why did that work?," and then you're kind of going, "I don't know, but it works." It's very hard to turn that into an invoice is what was said last year. We kind of know roughly the direction we're meant to be working in.

Lastly, we talked about the rise of ... I think this is maybe a more gradual trend anyway, the rise of the specialisms within SEO. I think it's definitely something I've seen over the last decade of having people that did SEO, and now you've got roles for technical SEOs, and content, and outreach people.

Judith, Kristina, and Chris have kindly compiled some notes, which I'm going to attempt with shamefully little preparation now I hear how much preparation Judith puts into her show, to kind of pick through and just see where we go with it. What I wanted to start on, and I guess we'll start with you, Judith, is ... I think everyone had something about this. It's these Web Core Vitals. Web Core Vitals obviously was something that have been talked about a lot this year. I mean, what are your thoughts in general on Web Core Vitals performance and how big of a thing you think they're going to be? Because they're not a ranking factor yet. I think it's May, isn't it, next year where Google said that's coming into the mix.

JL: Yeah. I think with the Google Core Web Vitals, I feel like it's the perfect opportunity for any dev agency or organization to partner with an SEO and just target every single site on the planet. I was going to use a swear word. I'm trying to be better.

MC: You've used one now, so the floodgates are open. It's an adult episode now.

JL: It's adults only. That's where I used to work. No joke. It was a chain of porn video stores called Adults Only.

MC: Get back to Core Vitals, Judith

JL: That's another story, although those were some core vitals in there. Core Vitals for the web, I think one of the things I was asked ... It's funny. I went on a rant at somebody about Core Web Vitals and what should SEOs or marketing managers do, physically do, to prepare. I was like, "Well, they could take some ibuprofen and just calm themselves for the discussions with development, and also buy a lot of doughnuts because they're going to need to suck up to the devs a lot." Why? Because this is not an SEO task. Core Web Vitals are not an SEO task. We need to know about them because it's going to impact SEO. It's going to impact where we rank in a head to head battle. If you're number one and you drop to number two, it's a massive hit you take on clicks, so yeah, we need to know about this and we need to know that this is going to impact us, but Google have been on about speed for years now and people still haven't clued in.

I had a client the other day say, "Yeah, they said speed matters, but I don't think so." You know what? I'm a consultant for these reasons. I can walk away, and in May when you call me up and say, "Yeah, we dropped a lot of spaces. I don't know why. The site is still the same, but we dropped," I can be like, "Well, you remember the report I sent you." Yeah, it's so important. When that icon comes in based on your speed and your experience of the users on your site based around these Core Web Vitals, it's going to be so big on click through rate, and it doesn't matter if you're number one at that point. If you have an icon that says you have a * * * * user experience, then you are not getting the click. I don't care who you are or what you're doing. That icon, if it comes in and it tees people up to not have a good experience on your site, they're not going there. They're not going to take a risk.

JL: Our time is precious. I know in lockdown we have a lot more time, people say. I have never been this busy in my life because we're all going online and online businesses need SEOs. I think it's huge as far as the impact, but there's nothing that an SEO, unless they are a web dev and they are directly developing a site, can do.

MC: Chris or Kristina, what are your thoughts around speed in 2020 being a ranking factor? From my experience with clients, I feel I've always said speed is secondarily an SEO thing. You know, performance. It's like we should be doing it because it makes users happy, and there's more conversions, and it affects the bottom line directly. It's like a bonus in that it helps SEO. I feel that the moment before Core Web Vitals, if your site was two seconds or four seconds loading, it didn't seem to make a big difference, but there was a difference if it was really slow or really fast. Do you think Google is now ... Do you think we've previously seen Google undersold the importance of how important it is as a ranking factor and that will change?

CG: I've always found that SERPs correlate to the quicker sites positively. If I work in any niche with a painfully slow website, we always look at ignore what the absolute score is. What is it relative to the other people that they're going to click on? Who are they going to bounce and go back to? Usually ... I've got to be really careful here. You're not going to get a really terrible site at the top if everyone is in the same niche in a comparable space. If you're an eCom and you've got loads of heavy Magento sites, or worse, kind of sat there, they're all going to be broadly as crap as each other, but in that sense you're still going to be the least crap. I think there's a degree.

The thing this year that I've stopped doing, and I've labored the point, is if the only reason that you're saying page speed is important is because of SEO, then that argument is flawed and wrong anyway from the start. It needs to come from eCommerce, the conversion point of view, and I think that needed to change long before Web Core Vitals.

The only thing I'm slightly skeptical of is the impact of that when it hits in May. Looking at when ACTPS became a ranking factor and then when Mobile-geddon happened, it's not that I don't think those are important. I think they're all incredibly important and you should have been doing it all along, but I think Google is going to chicken out a little bit in its rollout. I think the difference between the mobile indexing and how long it took to properly roll that out, I mean it was years, wasn't it? I think I still come across a handful of sites that still aren't and we panic, we moved a lot. It's kind of had vague memories of the millennium bug in so many ways, which is maybe a bit more of a nostalgia, but still it's important. It's got to be.

I was on a call yesterday and someone said, "Oh, Web Core Vitals is big for us next year," and I said, "That's cool. What does your dev pipeline look like in May?" He goes, "Oh, we don't have any dev support until May," and I'm like, "Then you won't get any Web Core Vitals fixed between now and then, in the short term, unless we can stick you on Cloudflare or similar and run through some quick fix optimizations." As Judith said, it's a dev task. Everyone was like, "Well, it's an SEO issue." I was like, "I wish it was," because most things are a dev issues.

MC: I like this idea of this prey theory of just being faster than your competitors, like we'll run away from the lion. I don't need to be this fast. I just need to be faster than you.

CG: Be the least * * * * , yeah.

MC: Kristina, in terms of eCommerce platforms, I know we've spoken about stuff like Shopify and WooCommerce. I know you worked with Magento, as you contributed to that Content King guide. We had Mordy on the other week who spoke about Wix. Have you seen much in terms of improvement or focus of these platforms on performance? Is that something you're talking to eCommerce clients a lot about and is there any particular platform you see that's kind of preparing well?

KA: Yeah, so I wouldn't say that I have 20 different clients which have different platforms so that I have enough data to support this platform is doing this, this platform is doing this, so it's hard to say it like that because I need to have real data. But I'd say that, as Chris pointed out, Magento is still slow and it's a huge beast. I used to work for five years actually for a company that created Magento extension, so at that time I was really into Magento's sphere. I know that it's such a huge beast and there's so many moving parts, so it will take lots of time for them to change this.

Shopify, they're empowering their SEO team. They are doing so much now, but also what's important is that speed is influenced by themes, and Shopify doesn't control themes. It happened to one of my clients. It was actually on WordPress, but the things with the theme, they has a small image. I don't even remember. It was a logo or some part of the page, but in reality it was a 4K image, which was loaded every time a page was loading, and only a small portion of this image was used. No matter what content management systems are doing, if themes which are developed by people, which I find these teams don't have SEOs for some reason, I find it very often. Why are you doing this? But as long as it's happening, no matter what content management systems are doing it will not improve as much as it could potentially. That's why basically you need your own developers and you need your own SEO consultants.

Also, to add to Chris' point that it's not going to rollout that quickly, I totally agree because we remember exactly with mobile, mobile first indexing the year of 2016 or something. I came to my boss and said, "We need to make our website mobile friendly. It's a thing. We need to do this right now," and we did it. In 2021, our internet, like Google, is going to be mobile first 100%. It took so many years, so I think it's going to be the same with Core Web Vitals as well.

JL: Let's hope. My gosh, can you imagine all of those people?

KA: Well, yeah, and then May 1st all websites drop and now only you or something. No, I don't that is going to happen, but we'll see.

JL: Yeah. As you know as well, speed is important for conversions.

KA: Yeah, like definitely, 100%. That's why if we have something like image optimization, this is the easiest thing to do, but at least that's something to start with.

MC: I feel like I have to dive and take a bullet for Magento here. You can deploy, just to give them credit, Magento now headlessly, so you can do your own front end, but you need some dev muscle to pull that off. We've got a client at the moment that does this, so it is this kind of heavy weight back end, but they have actually managed to get the front end pretty quick now. When there's a will/competent developers, there's a way. Chris, I know you're a big user of Cloudflare because didn't Shopify ... Was it this year or last year they kind of moved a lot of their infrastructure onto Cloudflare and now they're in this ... What do they call it, gray Cloudflare? You can't put your site properly fully behind Cloudflare, but their whole infrastructure is. Is that correct?

CG: I think you can go Orange Cloud on much more things than you could, so initially Shopify's relationship with Cloudflare made it more complex because you couldn't play with all the cool toys because the mothership was determining that, but I think it's much stronger than it used to be. I think actually for things like Cloudflare, even AMP and similar, there's a lot more out-of-the-box stuff that's going to get people much further quicker and it removes the human error bit. That's the one unifying factor or course in any CMS we talk about. The platform can be amazing, but if the person building it is not at all competent, I'm being polite, then you're going to still suffer. If you've still got a two meg image that's shrunk to a 200 pixel square, it doesn't matter what infrastructure you've got behind it.

I had quite a long talk with Jono Alderson about this, and we come to the firm conclusion that maybe we shouldn't trust everyone to do the right thing here. I mean, for those that have the will, and the way, and the investment to do it, for those who optimize for Web Core Vitals before May and can get ahead of the curve, they're going to be laughing all the way into 2022, but I just don't think those lagging behind are going to be suffering as much. Christ, I still audit websites that are on HTTP and they're still ranking in the mix. That said, I never have clients who aren't ranking absolutely anywhere because usually they're out of business by that point, so I guess there's a bit of selection bias around that. We will always see the oddities, but those who run and embrace Cloudflare and a lot more of the edge based stuff, or even the optimization on the edge like taking the pain out of images, webp and all of those other kind of new formats. Why wouldn't you click two buttons and have it all done for you rather than have to configure it on the server? I don't know. Fight me.

MC: While we're on the subject of doing everything right but sometimes not getting the payback from that, did anyone suffer at the hands of Google's indexing problems this year?

KA: Oh, yes, yes, yes.

MC: Terribly.

KA: Yes. Oh my god, when the request indexing disappeared, I was like, "Oh, okay." When you have a new page appearing on the website, it's not like you always need to push this button to tell Google, "Hey, hey, I have this page," but then something happened. One of my clients, I accidentally noticed that they for some reason had robots.txt working the whole website, and it's a client that it's a website of their client. I noticed that it was happening. I urgently emailed them and said, "Oh my god, you need to fix this right now, ASAP." They did it. I said, "If you fix this right now, it means that the chance is high that nothing will be influenced badly. The client won't see it and it's going to be okay."

KA: They fixed it, and then five days later the client is like, "Our traffic is dropping. What's going on?" The reason is that Google quickly ... Thank you, Google. Picked up robots.txt and they even for some reason put canonical tag to the non-HTTPS version, which was also under robots.txt. Then when Google finally picked up the right robots.txt, they still didn't re-index the page, which was still canonicalising the HTTP version. It was a mess. I was checking all the time. I was like, "Oh my god, how do I make Google know I need the cross-indexing right now?" Actually it was the hardest part of this thing for me because before that there wasn't a time when I desperately needed index file, but it was one of those times when you need it and it doesn't exist. I wish it appeared again, but it can be more. You can ask more questions, like not to let anyone request indexing with any page. There can be restrictions, but not disappearing at all. No, don't do this. Please come back.

JL: I can understand why that would freak you out.

MC: Have any of you noticed pages becoming indexed slower? Because I know that's a problem.

KA: Yes.

CG: Yeah.

MC: Is that something you've noticed, Judith, with clients, or Chris?

JL: I've got a publisher client, so I can say that every day I go to look at discovered or not yet crawled, and so it's not even indexed. It's not crawled and not indexed. It's discovered and not crawled, so it's the next level up of screw you, and so I have seen this. It does really impact us because those are key articles and they're just not in the index that is returned in search yet.

KA: Oh, and there is freshness, but there is no freshness anymore. You publish it. You just indexed it. We published it a week ago, right? Yeah, I notice the same with my own website. It used to re-index really quickly, but I posted an article and it took like a week to index it. It was crawled, but it wasn't indexed. It's not like it was impacting me, but even with these minor things I noticed it.

CG: Yeah, the issue we've found is, if it's JavaScript reliant as a page, the indexing of that feels much more painful. That's obviously only a handful of cases. The workarounds we've been trying to get to is, if there's URLs that we want focused on, stick them in a new site map and submit the new site map, but whether that's akin to doing a rain dance and hope. Then if you notice it then rain, you're assuming that your magical dance has worked or not. I think that's kind of in the territory, so there's been a sort of a shuffle or a degree of a do what we can. Even in some instances, do we change the response code of the page? Do we put it to a 503 or similar and try and trick Google into revisiting and coming back? But I wouldn't advocate any of those methods as a hard and fast way of doing it. Again, it's our version of a rain dance to that end.

CG: What I will say is external signals are probably the biggest variable in that. If you do publish it and you note a link, or you get a lot of buzz on social, they seem to be working better now. If you're publishing hundreds of articles a day, or a week, or whatever it looks like, then 98 of them are still buggered, but it's still kind of rollout, push the content hard, and it still kind of gets in. But Google has been drinking this year, haven't they? There's a degree of what's going on.

MC: We've covered in the podcast, I think it's four indexing issues they've had over the last 20 months or so, and this is just ones that the public have noticed because I'm sure they fix things obviously all the time that don't get noticed downstream.

KA: Yeah, sure.

MC: Four times where things have been bad enough where people have said, "Hey, we think something's broken."

JL: The last three months I think we've seen two or three major problems with the index that were not quickly solved. It took hours, and in once case was it weeks? That one where they lost a chunk of the index, so yeah-

MC: This was like in a point something percent of the entire index as well, so this is a huge amount of pages. It sounds tiny as a percentage, but you lose 0.1% of the web, the visible web. It's more than a blog spot.

JL: And it wasn't a fast fix. There was something seriously wrong. There was something structurally wrong, and that's just what we've seen in the last few months. Then of course we had the loss of everything when you were logged in. All of a sudden nothing worked, so something is afoot at Google. What it is, it's probably structural. It's probably networking stuff, but there is definitely something rotten in the state of Denmark and it's only getting worse.

MC: So this is something we spoke about and I kind of had my theory about, which was in Google's podcast they had spoken a little bit about the crawling and index, or mainly I guess the ranking issue they had. When Gary spoke about that, he talked about crawling overwhelming the indexing and creating a bottleneck, so the crawler was dumping too much stuff on the indexing. This is completely my wild speculation, but then a few months later I feel like I'm seeing a lot more pages getting crawled and then not getting indexed for days and day. It made sense to me that maybe something has been even temporarily adjusted here just so they don't get that big backlog.

Google of course disagreed with me when I put this out publicly, and I said, "Maybe this is related to why they've got rid of this request indexing at the moment," because it seems to me logical that request indexing kind of skips you ahead of that queue of you've been crawled and this is how important we think you are. You get in the queue and we'll index you when we can, whereas the request indexing infrastructure would just be like, yeah, these get to the front of the queue. The amount of people using that obviously is massively going to increase if this happens. Do we think there is maybe something to do with the mobile indexes? Is there something, do you think, Chris, Kristina, actually kind of rotten, as Judith says it has?

CG: I don't know. I can well believe that certain parts of the infrastructure outpace others, especially if how they're indexing pages, going to Evergreen version of Chrome, and sort of that whole portion is a massive technical challenge in its own right. If they're just a little bit slow, like half a percent slower than they were previously and it's a finely tuned system, that's going to have an overlap. What's probably worse though is, if people are noticing indexing is taking longer, more and more people are going to use the manual submissions. I mean, granted we as a population of people are relatively smaller than web users, so I'm not sure that the SEO community is going to give them too much grief, but equally if people are at scale hitting the request or index-me-now kind of button, that's only going to make matters worse at key times. I wouldn't be surprised if we don't see it coming back, certainly not in the short term, because I never-

JL: Bloggers.

CG: I never really understood why Google would give people the option to do that. I mean, I like it, it's incredibly useful, but it-

MC: That's why, because it's incredibly useful.

CG: But why does Google need to make our lives easier? I don't know. I mean, I want it to because we are the naughty toddlers, aren't we, of the internet. Do you enable the naughty toddler or do you discipline? It's a tricky question, right?

JL: Let them figure it out.

MC: I want to move on from ... I don't want to get stuck. I mean, maybe we can have another discussion sometime about the rendering queue and the disparity between Google saying that in almost all cases JavaScript is rendered before pages are indexed. I would love to talk about that, but I think maybe that would go on for the rest of this episode. We've definitely dealt with some indexing issues in 2020. I guess the last thing I want to talk about is a couple of you have made notes about, I guess, the fragility of search results in terms of two things, which is the broad core update, as it's very helpfully called, and actually this shift in intent with COVID as well. Chris, you've put here in your notes about ranking number ones can overnight suddenly mean nothing if the intent changes, right? Do you want to talk maybe, or Kristina, about the kind of impact that had on people you're working with, this change of intent and what we've done to try and help people around it?

CG: For a couple of brands that we work with that actually dine off their brand equity, if the intent is gone you can still be ranking firmly at the top for your own searches, but obviously the traffic has been dropping massively. When organic traffic drops, that becomes the SEO department's problem, but if the SEO department are like, "Well, we're still here." I'm not going to take a sandwich board out onto the street and start yelling that people need to search. For everyone that’s really suffered for that, there are others that are the beneficiaries of that because equally we've had people that have been ranking six or seven outside of genuine traffic, but that uplift in the intent to purchase online equally ... You know, issues with fulfillment and stock. You could be ranked at the bottom of page on or even top of page two and you could still see record sales just because of that radical shift in terms of intent and how people actually purchase. 2020 will be the year that renders year on year comparisons were rubbish. Next year's report is going to have-

CG: When we've got clients that this year have said their April was bigger than last Christmas, that means next April is going to look pretty trashy by comparison assuming that we're not in exactly the same ... Now, equally you do assume that businesses have stakeholders that can rationalize that kind of activity. There's a mixture of nodding and shaking heads going on here, isn't there?

MC: I'm nodding to you'd hope, but ...

CG: So it is a massive difference. I've said this more than once recently. Okay, if intent is down, you still want to be ranking for when intent comes back, so it doesn't actually change the overall dichotomy that you have to be visible for when people are there. It's just for when people aren't there.

JL: Yeah, and also stock levels are incredibly important, so making sure your schema is up to date so that you can prove in the search results it is in stock. I have a client who is more expensive than other people, but because he is able to demonstrate that his product is in stock when other people it's out of stock or it doesn't show what the status is, he gets a click, he makes a sale. It's like you say. It's the intent. Is that the same intent that it was before? When I type in, I don't know, Horrible Histories, am I wanting the TV show because that started recently? Do I want the board game? Do I want the books? What is my intent when I'm typing that in? I don't know. It could be anything. He's got nothing to do with the TV show. He's got the board game, but it's difficult if you're not there. If you're there at number one, but people want TV, your clicks go down. If you're there number one and when that intent shifts from the TV program to closer to Christmas, any products related to the TV program, then boom, you're making the sales. If you can prove that it's in stock with schema, you're much better off than a lot of other people.

CG: Don’t we intersect with the indexing issue as well though. If it were relying on schema for the in stock status and Google is not indexing.

KA: Yeah, that's a big problem. I also noticed when Covid hit, I also remember this really, really great research done by Lily Ray where she also showed that shifts in user intent and where websites which were more offline based, for example classes or something like that. When you're looking for, I don't know, learn SEO, sometimes you would see ... Before COVID, you can see real classes, offline classes that offer SEO courses, but now it's mostly online. It doesn't make sense for Google to show offline courses, so if you provide them the chances are high that you will fall in the search results, but at the same time is it a problem for you? No, because your classes are closed. You are not providing them anymore, so while you can't have those ranking it doesn't mean that you should at this point, like for these particular examples, when you are offline and people want something online.

Also ... I forgot the second point, but I really loved what Judith said about structured data. Show and use as much as you can. In Google My Business there's so many different opportunities to also show that you, for example, have online appointments. Utilize all these possibilities to provide as much information as you can before even a customer opens your door. Help them open this door by providing this information first.

Do we feel now going into 2021 that as an industry, and by that I would define people using search to purchase products and services that they want, do we feel net we're going to be benefited by this because essentially lots of people have been forced to do this during COVID that maybe didn't do it before and now they've realized? My parents never used to do online shopping, right? They used to be like, "Oh, I like going to look at the potato I'm going to buy," and I was always just like, "You're mad. You're wasting your time. It'll be fine. Just order them online." Now they've started finally ordering online because they didn't want to go out, or they couldn't, and they're like, "Oh yeah, this is really easy." Do we think that's something actually, as we come out of the pandemic, we're going to see maybe an uptick in spending online now people are...?

CG: Oh, yeah.

MC: Do we think people are more comfortable with it?

KA: Yeah.

MC: Or are we going to see the opposite? Do we think people are like, "I'm fed up with doing everything online. I'm going to go into every shop and buy everything, and touch it, and feel it, and breathe on people."

KA: I think it will be at first people might, might think like, "Oh my god, oh my god, I need to do everything offline," but then there will be balance. But honestly people usually prefer comfort over anything else. If they find that ordering something online brings them more comfort, even if they can go somewhere, but it sames time, it's more comfortable, they will definitely do this. I feel, I know that this shift to online shopping will persist. The percentage might be lower in the end. I hope the end is soon. I mean, the good end, the end of COVID, but these dynamics will still persist.

MC: Right. Let's look ahead into 2021 because I realize time is going really quickly here. I don't want people to start looking at their watches and start edging away, so let's think about 2021. I want to start with I think these broad core updates we've had. I just want to touch a little bit more on them because are we expecting? Do we think we're going to get more of these very creatively named monthly broad core updates?

KA: No.

MC: Yeah, no animals anymore, and then we're going to have the same tweet three times a year that we're meant to read about advice of core updates. Firstly, I guess let's start with Chris. What do you think about the guidance Google is now giving on these core updates? Because it's copy and paste every time, right? Is that helpful?

CG: I think the guidance is based around the idealism of what they're aiming for. I don't believe it's based on what necessarily reflects every individual change and I think there's a degree of actually the less information you provide around each update, that it's almost clearer and it's simpler I guess. The minute you say, "We've done this, we've targeted this and that," it helps people recover quickly, but equally it helps people take the piss basically and actually game the system. The more you understand about a change, the more it can be reverse engineered.

CG: I'd say that part of this problem and part of the ambiguity around it is because what they're doing now is so much more organic in the sense of what signals they collect, how they weigh up the signals, and actually the complexity of what evaluation is taking place. I don't actually think they can provide any more specificity. I think one of the things, this yo-yo-ing we see between updates, there's constantly this line that, no, it's not a reversion of the previous update, but when you look at the visibility trend of even someone like the Guardian, they are in a downward trend over the last three or four years, but there are some massive drops, recoveries, drops, recoveries albeit on that downward incline. Whilst the general travel is the same, they're obviously feeling their way ahead I would say.

You can't look at any one domain and make that judgment because equally the team and the infrastructure on that website could be part of the problem, but there is a learning by doing factor. As it's becoming more complicated and as the evaluation is there, John Miller was on Twitter the other day saying, "You guys worry about links too much." There's a playful nature to that and you won't ever convince some people in this industry that links aren't the most important factor, maybe myself included too, but there's some truth in that statement thought, isn't there? That actually maybe we are relying on other stuff that's just ... Sorry, maybe we're just ignoring the other stuff, but they don't want to give voice to that in any specific sense because, again, why would they?

If they can get the industry optimizing towards where they're moving to, there's a bit of utopian kind of thought around that. Whether or not you need to wake up and smell the coffee a little bit in between, I don't know, but we're not going to get any more specificity. I think they're going to between three and seven months in between depending on the size and magnitude of it, and I think it will almost expect ... You can't really judge an update until they're three or four weeks post because that's when you get the dial backs. You get it rolling across all the different indices, all the data sectors, and then people are realizing maybe we need to tweak some of the criteria and filters again, which you should be expecting, but it's complex.

MC: Judith here in her notes put something about this core update, saying essentially the fundamentals have never changed. Serve the needs of the searcher clearly or die. I think that-

KA: I love it.

MC: I think that's sums well, but I guess my question to this ... Because this locks in with what Chris said about this kind of utopia of just do this, and serve the user, and everything will be fine. Do you think this is serving the businesses well and how do you think this is going to go in 2021? Because we can't look aside from Google has taken big bites out of recruitment industry, flights, hotels, anything where for the user they can provide the instant answer and, hey, we don't need your website anymore. Just give us the data. How do you think this is going to go in 2021? Is there a way to fight against what Google is doing or do you have to just go downstream with them and just adapt?

JL: Yeah, unlike jobs, flights, they take a cut. With jobs you don't have to pay to be in the job search result and they're not getting any affiliate income, but in flights, the last time I checked they were still pulling in some cash on your choice to use Google instead of going to the website direct or to another flight comparison website. I do think that Google needs to really clearly pull back from that income model because, if they're going to do that, then we're all screwed as they successfully take things.

They were looking at going into food delivery, so Just Eat, SkipTheDishes in Canada, obviously I'm Canadian, Deliveroo, et cetera. All those places that were basically super-affiliates of various different restaurants, they're screwed if Google then puts that as something in the search results. Google is increasingly eating other peoples' lunch and they're getting paid for it, so it's not necessarily PPC. It's affiliate income, but jobs was the first place where we saw. You put job schema in, you get into a nice search result, and when you click through they actually give you all the different recruitment websites options. You can go to recruitment site A, B, or C as well as the poster's own. If a company like, I don't know, Taylor Wimpey or something like that has a job up, you could go there without them having to pay Google. I like that and I think if we stick to core fundamentals so to speak of optimization, optimizing speed, optimizing code, optimizing images, optimizing title tags so that people want to click, that machine learning element that is a big part of what Google is doing ...

JL: As Chris said, Google wants us to move in a particular direction. Why? Because there's so much more machine learning going on. It needs us to be moving forward, which is also why I said visual search is hot and to make images that are understandable by a machine learning element. We're all having fun going to Toonify and all those things that are making renaissance pictures of ourselves, or...

MC: I'm not, if you've seen my picture.

CG: That was poetry.

MC: I'll have to show you.

JL: But those are AIs. Well, they're machine learning elements that are understanding what a face should look like, and then deriving other elements and building faces. If your website has clearly identifiable products, then Google is likely to favor that over something where it's not a clearly identifiable product in the future. I'm not saying that's going to happen in 2021. I think this is a little further out, but Google's machine learning element is solidly in place, as is Bing's, and I think that for 2021 we need to start the process of understanding that and get our core fundamentals right. But that also means getting the images right, optimized, and well done, not just taking something off of Unsplash and chucking it up on the website. Core fundamentals, the stuff that we've been talking about since 2001, if not 1996, that is still important, and it's still relevant, and I wish people would do it instead of looking for the hokey kind of flavor of the month off of certain people whose initials are NP and what they're espousing of the moment. Fundamentals are still important, and if you get that right, then you can build on top of it.

MC: I think fundamentals are almost a harder sell to businesses because they take time, they're not particularly flashy, and when people are going around talking about growth hacks, here's a trick. Like you said, those things don't last the test of time. Chris, we spoke previously actually. I don't know if you remember. I think it might have been on Twitter about actually images and Google applying AI to understanding what's in images because they've got various APIs. You upload a picture of your dog, and rather than just saying dog it'll be like, oh, it's a golden retriever, really specific. Google gave some cryptic answers, didn't they, to “we don't use that AI in web search”. I think that's how they answered it.

CG: Yeah, it was about 18 months ago that I asked the question, and the answer was a very specific “it's not used within text search ranking”, but what was conspicuously absent in that was image search. I think the point that I was looking at at the time is you run an image through the cloud vision API test page and it will tell you, or it used to anyway. I don't think it does now. It would tell you which stock image sites did it find you on, or find that image on as well.

CG: More importantly, it could tell whether that image was unique to that, and then on top was that person happy, did they have a beard, what color was their shirt, et cetera. Yeah, why would you not use that information? It's just going to be a cost and processing ability, but for image search to work correctly they need to see it. If nothing else, just to remove all the obscene images people are going to start trying to plow into the system, but beyond that. If it's going to be a search engine that you want people to use, it's got to be effective and all of the debacles around their image machine learning and not testing it on the right size data sets and stuff, they've had the data now. They've been running it through. They've been training for a long time, and I think they're close, if not there already. I mean, wild speculation, but they're probably there. Must be, right?

JL: Yeah, look at the difference between GPT-2 and GPT 3, and also please don't try to use auto-machine learning generated text to rank because, if a computer can make it, a computer can recognize it. Just because a human can't doesn't mean a computer can't, but look at the logarithmic jumps between GPT-1, 2, and 3. You can see that machine learning is coming along leaps and bounds. It's always been … Not always. In the last few years it has been a part of the rankings, so we need to be thinking that way.

MC: I think that's a really interesting point from both sides of the fence there on AI. I have actually a blog post on my blog that's written by GPT-3 just to see if I could get it to rank for anything because we saw some really nice examples from ... It was Stacey McNaught who had found some webpages ranking for COVID masks, and the text they has used was the most basic, late '90s, like Marchov chaining or just the worst possible. No, I think it was just a mad lib in fact, thinking back, where they just switched out words to similar words. Obviously it made no sense, and it was ranking.

MC: I was thinking about, because a few people had said to me, because I had put the question out there. What do we do when we get past GPT-3, and what's the point of humans writing content? Because GPT-3 does a better job than a lot of copywriters I've seen, or people that claim they're copywriters. I won't say professional copywriters of course. It was just interesting seeing that if Google is still ranking this content that's just been mad-libbed and is awful, surely they're going to rank this GPT-3 stuff because new sites are using automated content already, right? When we get breaking news around stock prices changing, we've got news tickers hooked up to Richter scales so it know if there's been a 5.2 magnitude earthquake in this location. It can just get that news story out there. I don't know where I'm going with this because there's so many questions.

JL: Well, mad lib has been written and then is spun, right?

MC: Yeah.

JL: So it's not computer generated and then re-spun. Actually that would be an interesting test. It's to take some GPT-3 test, mad lib it, and see which of the versions ranks, but we're taking ... Or at least this is the way I do it. It's I take content that has been written by a human being and then mad lib it. That means that the patterns that you see in computer generated content aren't present. Even though you're mad-libbing it, the patterns of words and language usage is broken by a human and then mad-libbing it doesn't' change that. It's just * * * * , but you're using particular spun ... I really shouldn't be telling people how to do this. Particular spun words that are going to change 85% of your content, but the meaning is retained without making that, so it's a pattern recognition thing.

Same thing when you look at a YouTube video. YouTube uses machine learning to figure out if the video is the same or not, which is why you get ones with snow, because it breaks the algorithm. Mad lib stuff, if it's written by a human, breaks the algorithm, but GPT-3 has still got that ... There's a particular patterning of word and language usage that is recognizable by a computer, so theoretically Google should be able to tell GPT-3 text better than a mad-libbed piece of text because the mad- libbed piece of text has to be rendered and comprehended in a different way than the GPT-3, because it can spot already that particular pattern in the GPT-3 text that isn't there in the mad-libbed text.

CG: Interesting.

MC: That is an excellent point. I have a follow ... Is that though a shortcoming almost in Google's reading of the content? Because you said obviously if a human reads mad- libbed content they're instantly just like, "This is crap." With GPT-3, you read it and it takes you normally a few minutes to be like, "Hang on a minute. That was weird, that sentence," so actually when text recognition gets better I guess we'll see mad lib maybe sharply drop and GPT-3 stays. It's not going to unlearn GPT-3, I guess, or that similar content.

JL: It's about semantically related words and understanding the context of the semantically related word within the sentence. If feet, and toes, and heel, and sole, and foot, et cetera, are all related, they're all words that sort of relate to the same part of the body, but mean different things. But they're still related. It starts to get very nuanced. Feet, and toes, and foot, and shoes, et cetera, can go on, and on, and on. There are all sorts of relationships with these words. When you're replacing them in a mad lib, not only do you as a computer have to process that, okay, this is a semantically related word, but it's not the right word in this sentence, you have to understand the language and how the language works.

You also have to understand the nuance of meaning within the sentence, so obviously going beyond BERT, but the nuance of the meaning within the overall sentence and the overall sentiment that's being expressed within that sentence. It is getting closer all the time, so bad mad-libbed content will start to drop off, but you've got to think about our brains process so much of that. Think about somebody who is not an English speaker, so something I've just been working with now with one of my publishers, the devs are in Spain. We spotted a title tag error with a missing word "the" in one iteration of some auto-generated title tags because in Spanish you didn't need the word "the" in that position. Google is fine with it because it's a fine sentence, but a human being isn't. Where a human being is able to process that and understand it, you need to catch up. Machine learning is never going to be AI in my lifetime. It's going to learn and get smarter, but there is a limit, an upper limit, and it's processing time.

CG: One of the things we're missing, I guess, as well to this and what makes content credible is the actual person writing it as well. You think of the preoccupation of authorship and real author, real publisher all those years ago, and that-

KA: Oh my god.

CG: That project and what it was trying to do, and obviously they canned it because people couldn't implement it, but equally they said, "Well, we're kind of getting good enough to do this without you messing it up for us broadly." It would be interesting to see if ... Say the example earlier of that well known blogger, initials NP, decides to start using GPT-3 to author content rather than borrowing others. It would be interesting. Does that still rank? Will the sophistication from Judith, from what you're saying ... Presumably we go, "This isn't new." I mean, NP. Ignore the fact all of the ghostwriting and borrowing of content that goes on. I think that horse has bolted, but you get the picture, I guess. We have journalists or people who have been penning their own work, and then suddenly stop or they start getting machines to assist. It's interesting to see what the implication of that would be really.

I still think next year we'll see it happening more. I think in terms of auto-generating descriptions for product on e-Comm, and titles, and even the potential of actually we can run ... Can we split test content using GPT-3, like small snippets? Not long form, but where there's less context to identify the issues. I still think we'll see that next year. I hope we'll see that next year.

MC: Totally agree on things like using the Microsoft AI to generate old tags. I think that makes perfect sense because it is a good, that what you get out the end is helpful for the user. It needs to just maybe be scanned over by a person, but there's no reason you should be doing repetitive tasks that a machine can do because that's kind of the point of having the computer, as long as the output is as good as what we're trying to get to.

We've talked about the content side of AI that we're producing. On Google's side in terms of AI, we know they're using machine learning it seems more and more to fine tune the signals. Chris, you mentioned about Google almost knowing less of the specifics about these updates, so it's about they're running models and being like, "Well we think this is better, but we don't ... Again, that's kind of the point of a lot of these models. It's that it's looking at long lists of variables that aren't necessarily possible for humans to do. Broadening out the question, do we think links are going to become less important in 2021 than they are now? Anyone can answer that.

JL: No way. There's no way they will be less important in 2021 than they are now. In fact, in eCommerce internal linking and entity relationships are so important. What is a women's blouse and how does that relate to women's slacks and things like that, so that when people are looking for office suit Google needs to understand that what we mean when we say office suit ... Now, we don't always mean ... Obviously none of us are in the office, or few are in the office these days, so we don't necessarily need a suit, but if it knows me because it does, because personalization.

Hello, people. Wake up and smell the coffee. Personalization, it knows, and this is also part of the machine learning advancements. It knows who I am. I'm female, I'm living in Oxfordshire, so if I said best women's suit, or women's suit for office, or something like that, it might send me to the next website because there's a next near me and it knows that by suit I mean trousers and a blazer. It could theoretically drop me in the place in next that has a selection of suits, or it might select blazers for me because I just looked as a whole bunch of trousers and bought some. Heaven forbid that should ever happen. That's a bit scary, but if it knows that I own lots of trousers, it might drop me at blazers instead of trousers. I think there's a long way to go, especially for eCommerce where it can auto-generate content, but if you don't have unique auto-generated content, you're still going to struggle. I think, Kristina, you've probably seen this a lot in eCommerce.

KA: Yeah.

JL: If you have duplicate content, it's just not going to rank anyway.

KA: Yeah, 10 different pages for one product with only one minor change could have been just like color swatches or color options. Yeah, and in terms of links, internal links, oh my god, they are important. They are going to be important. I always say this is something that you control. It's not something that you need to ask people, you need to pay people to link to you. No, it's you control this, so you need to make this work on your website. Think about that. Think about your blog, especially your blog. I feel I see so many companies, especially eCommerce companies, creating blog posts. They're amazing, but do they sell? Do they actually tie to the products that you have? Do they link to the products that you have? That's the question.

KA: When it comes to external links, they are still going to be important, but I hope that these guest post requests stop at some point because I'm so, so, so tired of them. Yeah, just don't sell * * * * * links. Don't buy * * * * * links, don't sell * * * * * links because in most of the cases they won't work. Even if they do, it is not something that you want to risk because at some point there is a history. It's not like you are good with that now, but what happens in a year? What happens after the next core update? That's the question also to ask yourself. It's better to be safe than sorry because if you're removed from Google it might mean that you lose business. Some links, like they're just not worth it.

CG: Interesting. MC: Chris, do you think in 2021 we ... I spoke a couple of weeks ago to Lily Ray about some wild speculation patents Google had about identifying individual's voices online and also other patents about identifying their writing style. Google could potentially identify that I've written something or its me actually speaking. Do you think we're going to see, I guess, a progression in terms of ranking away from webpages and more to ... I guess you'd class it as entities. I hate to group us like we're just some other type of commodity. Do you think it's going to go to people, to individual specialisms, to identifying rather than the importance is it's this webpage because it's on this website, but actually I don't care so much about which page it's on, but I know it's got this person on it and they're important? Do you think realistically we're going to get anywhere near there next year?

CG: I think we're already on the way there. I think that's something that Google has been working towards for a while. What I would say is that I think to really feel the benefit of it you have to have a pretty big profile already, and I think you need to be a trusted entity that Google can verify quite heavily, but as their vault of all knowledge becomes bigger, more comprehensive, and they gain trust in it then it makes sense. I think who you are, how accurate you are, how much buzz ... I mean, I think they will have every reason to kind of keep hold of that. What would be really interesting is to run some tests on that, and actually separate people from their domains, and try and move them elsewhere.

CG: Kind of obvious example, for me, is Rand moves away from Moz. Does Rand's voice follow him because he's social and his audience followed him, or because something else did? I suspect there's an element of both actually, even if he's got every reason for Google not to want Matt to follow him now in some respects. Yeah, it makes perfect sense. I think the domain or website as a thing is quite an outmoded concept, or will become and outmoded concept. We see the small signs of that by browsers just stopping showing you your path of where you are because actually the user doesn't care and obviously apps don't have website. A bigger is bigger than an app and a website, and I think moving into the entity's space, we are bigger than our domains. As an SEO, when my Chrome decides to stop showing me the file path, I get all indignant. I want to know the protocol. I want to know the sub domain. I get really annoyed, find the site, and put it back, but I'm in the smallest minority because actually no one gives toss. And why should we? If the browser is doing it, what it should do, then the search engine should put us in the right place. Even just that differentiation of going between a browser to an app, it will become so it won't matter. The author as an entity makes more sense in that world than it does in yesterday's world.

CG: Whether or not we'll see any meaningful different in '21, it's one of these concepts. I liken it to a bassist in a band. You only know there's a bassist in a band when it's doing a terrible job, which may be unfair. My brother played the bass, so I used to rip it out on him a little bit. In that sense that I think this kind of stuff you only notice the jarring bits and when it's wrong, so authorship went wrong because I can't remember ... You know, somebody who has been dead for 200 years and is suddenly credited to an article. You notice it then, but every time it gets it right totally unthinking and seamless kind of thing. That's the direction, and that's search. We were talking about Google dropping the ball a few times recently. We only know it when it does it badly and we give them a really hard time for it, but every day we don't go, "Search is hard. Well done, Google," mainly because they don't need our gratitude. I'm going off on one now, but you get what I mean.

JL: Think about the search on thing with passages and video passages, so fraggles as Cindy Krum named them. Google is understanding the content of a video, it understands the context of your search with regards to the content of a video, and it drops you at the moment, or will when it's released. Drop you at the moment in the video that relates to your search, so think about how complex something like that is. You have to understand the video, you have to understand the content of the video, the context of the search to the content of the video, and then drop the person at the right point in the video.

KA: Yeah, I think it's already happening, but based on the description that you have on the video, under the video, like at this time this tab is happening, at this time this tab is happening. It's easier. It means that Google doesn't go through the whole video and needs to understand what's going on. It's based on their description, but yeah, it's fascinating. I remember when I saw it for the first time. I was like, "Wow, what's going on?"

MC: I think we're going to wrap things up now, as we've busted through an hour of our normally half an hour podcast. It's a shame because I actually have got so many more questions for all of you. Maybe if we could have a 30 second summary from each of you on one thing that you think you will be or you think businesses should focus on in 2021 in SEOs. If we start maybe with Judith, then Kristina, then Chris.

JL: Sure. The one thing that I would love to see businesses, especially small businesses, focusing on in 2021 is getting the core fundamentals correct, so image optimization, text optimization, title tag optimization, code optimization. Just the fundamentals, the basics. Just get the basics right in 2021 and the rest will be easier and follow.

KA: Yeah, I agree, and to this I would add also, for example, for eCommerce websites specifically. Think about what other, for example, categories you can create based on what you already have because with COVID especially I've seen many websites really successful when they ... For example, stickers. When they have specifically COVID stickers. Yeah, a potential customer might understand that there is a stickers company and they can sell stickers for COVID, but you should specifically tell about that on your website. That's important as well. Fundamentals and plus communicate what you have, what you have to offer. Don't make people think and make many steps because then they will leave and go to someone else.

CG: Okay, I think for me I've been banging on about this for this year as well, but intent, intent, intent. Understand what your user needs at that moment. Stop racing towards the bottom of the funnel and actually understand the difference of intent for different types of content, different types of searches. But also attached to that is the recognition of what does actually good look like in that space anyway because too many have actually no idea if the content is any good or not. They get very indignant because their site is not ranking better, because it looks better, but actually fundamentally people don't care if it looks good if they aren't answering the question.

CG: I think actually really spending the time there and a lot of that is what I'd call pre-SEO concern because actually we don't know who our customers are. We don't know where they are, we don't know what moments in their life we're actually fixing and solving. Obviously the SEO swoops in because actually if you're not hitting that at keyword research and content planning, then you don't have any hope. Usually from an SEO perspective how do we do this different, we ask that difficult question then. Who are your different personas, avatars, users, whatever you want to call them, and then how do they differ? Are they going to be on their mobile in the supermarket, or they sat at their desk at work? What's your context for your search and what are they likely to do? You can't cater for all of it, but you can damn well try. Usually the benefit of that is you go upstream from the competitors and you get that bite of the cherry sooner. Also, if you want to play nice with your ads team, retargeting can then kick back in. Again, you're creating a fuller experience. The other thing I advocate is stop thinking in silos because our users don't operate in them, but that's not a very popular thought in the SEO space because we're defensive like that.

MC: Brilliant. Yeah, you heard it here. Judith says fundamentals for 2021. No glitzy NP growth hacks for her. Kristina, I think don't make me think is a really good fundamental that you've gone over there as well, so actually about thinking about what people are actually searching for and creating those categories in almost realtime. Chris, I guess bringing everyone together and not just making webpages and content, and actually thinking about it's not just a website. It's part of this more complex journey between sales, marketing, mobile, desktop, and where the user is. We have to end it there. I want all of you to come back individually because I've got notes here of other questions I want to ask you, that I haven't had time. I want to wish you all a very lovely Christmas and thank you all for joining us. I really appreciate it, guys. Thank you.

JL: Thank you for having me and all of us.

CG: Thank you for having us.

KA: Thank you.

CG: It's been fun.

MC: We will be back on Monday the 11th of January next because I'm going to take some time off over Christmas and I'm going to go slowly back into maybe the office. Probably not. Probably just back into a room at my house on the 4th of January and work myself up to recording the next episode of the podcast in 2021. If you have enjoyed it, please do subscribe, share it with a friend, or of course give us a link as long as you don't tell Google. Thank you very much.

CG: It's all incriminating from here on in, guys.

MC: It is, so careful what you say now. It's all being recorded and I will use it for blackmail.

More from the blog