 Do we do an interview now? Is this how we... Look, all I wanted to hear was please. Action! Oh no, he's watching. So impolite. Oh my god. Who are you and why? Why, I still haven't figured out, but I'm Martin Splitts and I have recently joined the ranks of DevRel for search and web ecosystem at Google. So what are you here to do at Google then? Google search wants to basically get out of your way. That's the idea. You just do the right thing for your users and we do the heavy lifting on the back end to make that happen. So is this basically where, you know, all longest time certain techniques were no go for developers because it would lower your chance of getting to the position in search where you wanted to end up? Well, that's still a thing. Like you can still try to gain the system and the game is going to punish you for it. Like we are pretty good at catching these things, but nonetheless there are a few things where we can improve the way that we help you like get stuff on search and get like the relevant users to your page and that's like a bunch of best practices and there's going to be talks on IO here from my co-workers Maria and John, John Miller and Maria Moeva. All right. They're going to tell you more about how to make like JavaScript and search better friends and in the background I'm basically like figuring out where JavaScript and search are not friends. I guess we'll link to those? Yes, we will. Available? So is this about making things like like single page apps? Yeah. We have one of the biggest changes recently where now the crawler basically is able to execute JavaScript. So single page apps suddenly became a proper first class citizen in search results. But then certain people figured out like, wait a minute, you say it executes JavaScript but does that, what does that mean? Like if my page takes 10 seconds of executing JavaScript, is that fine or what if I do this? It's not. It's never, that's never fine, that's never going to be, but like there's a few things and we are running a certain Chrome version. We have recently published, well recently I think like last year we published which Chrome version that is. We are a little behind there and we're working on catching up and for the time being we have to like put out some guidelines for people to fix the issues. And we're also like getting into how the frameworks are doing certain things and if we can help them do certain things better or if they're doing things fine already so that like people get a better. That's a pretty different angle because before we were doing like VR, AR things, right? Yeah. So that's like the other end of the spectrum. That's pretty much the other end of the spectrum. That's nonetheless super exciting because obviously you want like people to find your stuff and it's cool if you build VR things. Who would have thought? Right, I know. And in the future hopefully not too far out, we might actually also have a look at how we get like 3D content into search so that you can. Oh no, because I hate looking for things around the store. Don't, is that what it's going to be like? Like searching on the internet is going to be like walking in there. And it's like you have to walk over to someone and go, excuse me, I'm just looking at one aisle page two. I'm just looking for some socks, no, no, no, no. Where is the sock aisle in Google? But if you today, if you search for a product, you might already see like a price tag and where you can buy it and the little picture. But if the picture is not really helpful, then it would be nice to just be able to spin it around, no? That is true. That is true. Boom. And I'm not saying that's going to happen. I'm just saying like that's something that I personally would love to see. You can already be quartered out of context now. It's amazing. We can do that with CSS 3D already, right? You can just spin it around. It's the same thing on the other side, just reversed, right? That would work for socks. For socks that would be fine. Probably not help, but you know. Or CDs, like a picture of the CD is probably, yeah. You don't have to explain that for the younger viewers. Oh, right. Yeah, we used to store music so that we can listen to it on the go, eventually, and it was like little discs. Very compact. Yeah, very, very compact. The size apparently is chosen so that some like Beethoven something, something symphony fits on it right away. Like Philips designed it that way. Is that right? Yeah, yeah. Oh, is that why it's 64 minutes originally? Yeah, yeah. That's the reason why. The CD can hold one Beethoven. Yeah, like 64.7 something, like it's ridiculous, but that's where I come from. So are you saying they could have made it better and just didn't because it fits a Beethoven's good enough? Yeah, yeah, that's fine. So if people are wanting to do better at search on the web, right? What's step one is it to tweet you and ask nicely and you can go in, press the buttons and the thing goes to the top. No, none of us can do that and none of us will do that. We also don't like offer any preferential support. So even if you're a partner, you're not getting anything that other people are not getting. So the best thing is build it for your users. Use the tools that we are putting out there like Lighthouse, PageSpeed Insights, there's the mobile first testing tool, which gives you a bunch of insights. Search Console is a thing that you can verify for your search on page and then you get like a lot of insights on how you're doing in search as well. And you can see how Google is rendering it. We get feedback on that. Yeah. So this is a kind of way to see how like the Google crawler sees your site, right? Yes, that's like it's called fascist Google. It's probably going to be called something else soonish because we're working on a new version of the Search Console. We like to rename things. Hashtag branding. Hashtag branding. That make it difficult for people to find perfect excellence. OK. Especially in the Search Console, things like finding things are important, right? But like this, the mobile first testing tool that gives you also a view of how Google sees your page and shows you like JavaScript errors and stuff. So it's pretty cool actually. That seems helpful. Yeah, that's useful. Because beforehand... You said like the version is a little bit further back, so certainly API is my throw, don't throw unstable. Correct. And you basically beforehand all you had is like you go fascist Google and then it goes white page. Sorry. Great. And then deal with it. So what kind of cases, where is that going to happen? Is this like a case of JavaScript errors or... Yes, in case of JavaScript errors, if you use features that like if you if your site breaks, if you use certain features that are not supported by the slightly older version of Chrome that runs the Google... Teeny-mild, tiny little bit of deviation from it. It's relatively not that old, relative to the age of the planet. I mean, like the internet is 25 years old or something, right? So it's like, it's a glimpse. Nothing but a glimpse. So if you've got that kind of heavily JavaScript based page, you're seeing errors, these things just like nothing coming up in the Search Console, what's your next steps? How are you going to improve that? I mean, you would have the errors not the console. If there were errors, so you can... Now you have actual features like... But what if it's like, you know, core feature not supported? Yes. Go away. Right. Then what... Just babel it. Babel everything. So John talks about that in his session. There's like server-side rendering and there's something that John likes to call dynamic rendering where you like render certain bits and pieces and serve it when you see the Googlebot. I know you ace niffing is not a cool thing, but you know... That sounds like gaming... The search system. You said that was bad. I'm going to have to pull you up on that. Here's the thing, like, you should not do it to, like, use this because we're going to catch that. But if you serve us an exact copy of your content for the Googlebot, then you'll be fine. You're not going to be penalized or anything. But if it's an exact copy, then what's the point? The point being that you can, like, layer on additional functionality. Like, if you give us a static HTML version... I see, so we're still rendering, basically. Yes, exactly. There was one saying, but then on the Googlebot you don't inject with Java and do the magic stuff. If you use, like, I don't know, 2019.js, which is probably going to be coming out very soon and... Probably next year. Yeah, right, I would guess. Or i018.js, which is probably announced right now. Then you might run into problems with search and then you can go, like, OK, but I can, like, have a server-side process render out to some HTML so that we get the content of your document and the relations between the different documents that your application consists of. But that's good for performance in general, right? That's good for, like, real users as well. Yes, which is why we are fine with it, because, again, we don't try to build something specifically... It's a win-win. Yeah, we don't try to make you something specific for search. What we want is you make your website better for your users and make it relevant and make it high-quality and make good content. And then it's our task to get you found. But we don't do interviews, mate. We never have done interviews on two or three. We're suddenly starting to do interviews now. Yeah, well, it's going to be awkward if we don't interview the person that is sat. No, it's fine. You know, I've got a phone with me. You can just browse the Instagrams. We just, you know, chat about the usual stuff. Well, make this more interesting, like, up the bed here. I'm going to use your phone. Well, I'd say tweet this. I don't recommend. I don't recommend.