Alan Kent, Developer Advocate (ECommerce) at Google
Oh today. Are joined by a very special guest, on Kent. Developer. Advocate. At Google. Previously. Magento. Chief, architect. Allen, how are you she, was. A chief with, a deep voice. It. Does it does you do you got to say chief with a deep. Voice like that. So. So. How's it going so yes oh by, the way I first want to say on on behalf of the entire Magento. Community, we, miss you greatly and. And. Also a, hearty. Obviously it's been a while now but a hearty congratulations on. The move to Google but. We do miss you how's how's. Everything. Going with it with the transition, to being. To developer, advocating, and at, Google. And. Apologize, for the blower. In the background hopefully they'll be good some neighbors hopefully, they'll be finished in a minute or two oh it's, not too bad okay, cool, the. It's. Annoying. Interesting. So the first couple of bucks is really sort of landing. On the feet understanding, how Google works understanding. That it's just really one perfectly, unified, organization. No. It's not, right, hand always speaks. To the left hand yeah. So. Many people, there with, so many projects, going on that it's just impossible pretty, world to coordinate, and it just seems to be part of the Google culture as well as to know they they encourage, creativity, they encourage. Considering. Other ideas, and so no, they, frequently do have a couple of, alternative. Strategies, running in parallel just to see which one's going to work best. Some. They, are more leading-edge, than others and that's, just the way Google works really. Culture, of the different countries companies is that's. Just a Google culture right, right, how big. Is Google. Now in terms of number, of employees, and number. Of products, and services it's. I'm. Not you probably don't have the numbers to rattle, off off top of your head but I. Was. Looking around and. Going. To the Google homepage and, brought up the number of products and I think it's got like 300. Different icons, or something or other great. Giving. A lot when, yeah I didn't realize you know you bring up your Google menu item and oh yeah it's got that and you got them clicking the more apps and then oh another hundred, pops up and then more, apps again. Through. The icons and said which, one of these do I think could have some possible, relevance to a commerce business right. And, it, was a great percent. They. All should be used but you know like Google. Docs well you know that's a communication, mechanism that could be useful and so, you go through it all and you sort of say in. Terms of being related to a business there's a lot in fact. It. Probably was more like 75 percent but, I, don't maybe would be wow. Over-ambitious. Bravery. Stuff I don't know if that's a official. But it's certainly in the many many tens, of thousands, of for, brands path so, right when. I turned up for the onboarding. Session. Know sort of like I wonder how many people are going to be turning up to the onboarding session this week, and. It was sort of like the queues out the door. Way. Out the back can barely see the person at the front and this was happening every single week Wow. Like. Just yeah a number of people there and the sheer number of people they're onboarding, all the time is impressive, yeah, yeah. I remember. Reading. Somewhere, about, their. Onboarding, process, and how I can remember the details but it was just really impressive the degree, to which they fine-tuned. Their. Onboarding. So that they're making sure you're right, fit you're in the right department. You can move from department, you know you're getting evaluated. I can't, marry many of the details but I just, remember being really impressed, by how finely tuned their. Onboarding, process was they. Treat it seriously it, was almost like two weeks of. Classes. You could attend. And no it, was great I do, all sorts of things right, up front right. Old. You what you're not meant to talk about right up front and then basically said look we want to keep. Everybody informed but. That relies on people, being discreet right. I want to encourage an open environment. But. Some things are sensitive so it's so like trust, the people but. Yeah. Not. Too much yeah. And, so, you work out of the amount the amount of you office. Which. Has got to be an experience, is it, do. You alternate. Between massages. And and and. And different. Benefits. Of the office there every other day what's what's that like. So. It's, interesting so no I haven't got any massages really. To what. People want to you you get you get points, and so. You can build up points and get some initial points so yes I probably could go off and get a massage but, but. You, know I didn't, really come for the perks I came, because of, all.
The Things that Google is doing in the commerce space and to me it felt from the outside that little bit too disconnected, so my. Personal, passion is to try and help connect, the dots a bit more help connect the outside world you know they gave me an, option of what sort of role and the one they picked was developer, relations because. I actually enjoy, getting out there talking to people, and. As we start doing a lot more of that in the future I just had to get my feet grounded at, Google. First but. Talk, to the people out in the real world bring, that feedback back, to the internal teams, because. Some, of the teams know a lot about commerce that was actually surprised me a little bit but, then a lot of teams had, no idea and so. You know there was one team saying oh we got this great example of how to collect. Credit card information over the web and, it's sort of like. PCI. Conformance. Oh. Wow. And, so it was just a bad example they were using no just trying to be relevant to commerce but they didn't have the commerce debt so part of it is bringing. Different teams up to speed and so I sort. Of like one of the people who's part of the bridge between the outside world and the inside world, get. Feedback, from the outside world bring, it inside to the actual teams get them did changed directions if needed and my, versa that take the messages from the inside teams and help communicate, it out to the outside teams right, that's, really cool how is, it sort. Of it's, interesting because when you were at Magento as an, architect, as the chief, architect. You. Know that's. Obviously in an architect, level role and I remember always feeling like you were so engaged with, the. Community. Probably, more so than any. Other architect, I can think of that that comes to mind and so. I, mean there's, probably, some, element, to which you maybe had that hat on even. In an architect, role but how does it feel to. Be in a in a pure developer, relations role beast versus. You know more of an architect, role which it seems like throughout. Your career there's been more the. The role that you've been in yeah. Well I think architecture. Is a lot about communication like. Somebody's. Like, what's the best definition, of architecture. Hey. How about I ask you a question like in computer architecture how, would you define what architecture, is. How. Would I define what, architecture, is I. Guess, I would define, it as sort, of. Designing. The, way that, a system, would. Work and. You. Know balancing. Out with lots of different constraints, you're, making me really nervous if I wasn't nervous enough, Allen you're making me nervous here put me on the spot but I appreciate, it that's a pretty good definition so, my favorite, definition I've come across is architecture. Is design, that matters. The. Question is based on the context, what matters and. If it doesn't matter no, you're. Like oh does, a sort algorithm, matter which algorithm, you pick well. If, you've got ten items to sort know it, doesn't matter mmm. You've. Got potentially a million items there and it's fundamental to the user interface then, it does matter and so, to me architecture, is trying to work out what parts of the design matter. Versus. What you just leave to other people a, lot. Of it is the higher up you go is really about coordination, and communication because. You know the magenta team got lots of smart people who. Frequently, no matter about getting them all the work together so, there's a unified, message at the end of the day that it all fits together well at the end of the day, that's. Frequently what the higher level architecture is about but.
If You don't really understand, the points so, you don't get in there and get your hands on which. I often do behind the scenes like I tinker. I'm. Not very good at self-promotion sometimes. So I don't. Really care about getting up in front of lots of people and sort of saying oh yes I designed this now, I'm quite happy talking to people behind the scenes and I'd like seeing people grow so trying to talk to them make sure they know what's going on guiding. Them through the process letting. Them tackle. The problems to me is that's. How people grow I really. Enjoy that side of things this is possibly my own personal mentality, mm-hmm. I love. That that's that's actually, really interesting because it ties in very well, to one of the questions I got leading. Up to this from. Brian Fowler which was. Exactly. What you're saying like from looking at your blog posts, it seems like you would kind, of explore concepts, tinker with things. Throw. Ideas out there and, then like did, you sort of take those it was just sort of like what you just said did you sort of take those ideas pass, them off to other architects. And team leads to, sort, of implement them and flush them out, frequently. Yes and so frequently the tinkering, is just enough to work out is, this worth pursuing a graph QL you know somebody, brought up should we be know one. Of the major things about Magento two was getting all the api's in it, was it step towards better modularization. Separation. Of UI from back-end, you know for. Architecture, it's frequently going to have multi-year plans and so you just step through a bit by bit you can't do it all at once so, that was a big thing and, so we got REST API we, made sure the code was independent, of the actual protocol so we've got rest and soap and you know it's got a few different ways you can talk to it without any, change to the underlying business logic and so that was one of the goals and then, along comes know somebody else was a Eugene, and sort of said hey what about graph QL, ok. So, it's a matter of well, how do you evaluate it well frequently you build a little demo you put something in together and you work out how good the tools are because. It's not only how good the technology, it's how easy as it is to use and the, thing that really impressed me about graph, QL is it ticked, multiple, boxes, it. Ticked. Reducing. A number of network hops it's. Actually got junk whites nice interactive. Developer tools with autocomplete, just in a web browser just, so to help you build your queries up so, it's the whole ecosystem, around, it that's frequently that, made that, particular technology, it solved, real problems and it, was fairly nice to use great. But. It was I going to build it all well, no I built some proof of concepts, made sure it was possible worked out I'm just I don't understood, what some of the problems are and then yeah another team then takes it on and actually delivers its through to production because there is a difference between tinkering.
And Production code and, I'm. Not gonna be around to support the thing I'm, not gonna be around to put all the polish on it otherwise I can't go be off looking at something else so a freer I do, tinker. To, work out if it's any good or not and work, out and, if you touch it you get a much better idea of hey is this thing really going to work or not. Yeah. Yeah. I I, really. Love seeing, that I think everybody loves seeing you tinker with things and and and. Jump into things and hack on a project, here and there it was really. Cool. To see I I thought, I think that um myself. I probably. Am maybe, more in the tinkerer, category. Than the you, know really, hardcore production, level code a lot of times that the products and things that I built they work fine at a certain scale which which is totally, fine for me but I. Think, I have a little bit of that that tinker in me as well. Of. Course of course, yeah but it's more a matter of where should you spend a time and you. Know if, you're spending time building a production system you're not thinking about your business so, it's. I think, it's more a matter of it's, not necessarily a matter of skill it's a matter of what should you be doing. Right. Has. M2. Sort of gone in the direction that that you've expected, you, expected. You, based on the, the, things that you were putting in place like you said those multi-year, plans. You, obviously came in at a time where you, saw it, by the way I see it kind of almost righted, the ship a little bit but. How, have, you seen, the direction go since you've left. Yeah. It was no. Do. You know he sold handedly right the ship no no there's a whole team of people but, there was a eBay, at the time sort of, put. A number of people in place to try and write the ship. Because. You know yeah. It. Had been promised out live and it hadn't got shipped okay well let's get that fixed and, so then it was a matter of trying to decide well how much do you do now versus, how much do you do later and it, probably got blowing up bigger than it should have been which, was know one of their hesitancies, but it was also hard to. Fix some of the things and not do, a good enough job and so it was a lot of. Anguish. Going through trying to work out well how much do we fix versus get it out there yeah. And how much you put in place now versus, iterate on later and that that's just this, is hard there's no question about it the. Front. End stack was one that was already done before I got there for example they'd already made done, less choice. And it's not that at the time was the wrong choice but, it took so long to get him to out but it might have been the wrong choice by the time I got live, and. Then became the question of when do you try and fix it versus because he can't destabilize. The community, too much no some people complaining, that we're changing too, much all right that's a hard one you've got to revolve the platform, and if. Something's not right sometimes, you got to take, the hard, things. And just turn adjust. Them and so we, put a Lunger quite a bit of the platform, in place but, then there was a bit of pressure to get more features out and so we didn't keep pushing on the platform and as an architect I would have liked to keep pushing on the platform a bit more yeah. I think, like simplifying. The whole data base model it's fundamental. To me to fix a number of the performance, issues and that's, what the graph QL to me is also a good thing it gives a layer of abstraction to, the lower levels, which, has got the potential to hook into the database api's, and so forth which. You can then feed into your indexing, engine which is one. Of the problems of people have performance, with the product but. Yeah not. My problem anymore yes. It's, it's it's really interesting I've. Talked. To a few people recently about PWA, and graph QL and how it. Really kind of changes the game from, multiple, perspectives, not only what. You mentioned having great, performance. Benefits, and things like that but even from a developer, resource, perspective, like all, of a sudden if this thing really moves heavily, into, the PWA, side. Of things with a very well. Well. Decoupled. Very loosely coupled front-end. And back-end all, of a sudden now you can get all of these you know.
Front-end. JavaScript engineers. Right out of college which is a very sort of sexy and interesting, technology, all of a sudden you have access to that whole pool of talent, which you. Didn't, really necessarily you. Know a classic. Magento, front-end, developer is a little bit of a different, beast. And, so. It's yeah. There was to, be completely honest when PWA came up there was really, the two benefits one. Is. Hate pwace are cool they're to know much more interactive, you've, got a lot. More capabilities. Blindingly. Fast if. You're out to code right of course but, they've. Just in their own right it's a very interesting technology but. To me it was also as much about hey. We can rejigged the front-end stack and just pick, a more mainline. Approach. We've, got the api's in place now because, before. The api's in Magento one know, that there were api's but it wasn't there, wasn't a platform, feature it was just like you get this extra module, and sometimes. This extension, would update the core functionality, but not update, the API. Behavior, or like you can't have that it's, got to plug in it's got to be uniform everywhere yeah, so getting the API is in was. Actually a really useful precursor. To get it but then that's the next logical step I think, I've, written about in a blog post before but I believe. That there will be more and more diversity. In, integrations. And front, ends you know you've got things like no voice. And Google assistant. You've. Lots. Of different, interaction, technologies, and I don't see how that's going to slow down I think it's something gonna get faster. Yeah decoupling, more, and more no Magento to put a lot of work into decoupling, them there was too much business logic in the, P HTML, template files in Magento one pulling. That all out separated, so now okay we've got ap is that give you access to the full functionality, and it, means you can cut the head off and. So people were successfully. Doing and have built quite a few headless projects, but. Then this is the next logical step now we've got to that stage well great now we can pick a modern front-end stack. And the, tools there is was as much of a. Excitement. To me around the PWA work as the, actual. Technology itself. Yeah. I thought, it was you, you talked to you, mentioned briefly, how. There's. The the blog posts you had written, on I, did a little bit of a deep dive trying to get ready for this on. The. Different you know technologies. In sort of media that there are there's vr there's AR there's mobile. Web voice. And, I. Think, your blog post, the. Future on the future of commerce one, of the ideas you had in there thought was really interesting was this idea that. Maybe. The future isn't so much as, a merchant. Let stick every single box let's get our AR let's get our VR let's get our mobile app and our, website but more so. Really. Understanding, what do you represent as, a brand, and what. Are the technologies, in the media that makes sense for you rui Kea, okay maybe AR makes a ton of sense, invest. Heavily in that because that's going to be a. You. Know an expensive, thing to really get right most likely I thought. That I thought that idea you was really interesting, yeah. And I'm not the only one saying it by the way but but. Yeah I truly, believe that if you just do should, you have a chat thought well, wrong, question. What's. Your business problem, where, are you not. Satisfying, customer, needs, where are you not getting cost efficiencies. In your business and then, pick a technology, that can best solve that problem so no, furniture, well yeah it, turns out to be a really great experience to, drop furniture, and try it out in your home room, that's, a real problem, people have like no I if, they go off the one to get a new sofa well. It looked good in the room will it fit in the room well. I can go to after with a, Nike, a store or whatever and I look at the furniture there but I still don't know how it's gonna really feel in my room and I sketch out a floor plan and roughly, sort of know the right measurements, but. So. That, was that's addressing, a real problem. Should. They have a chatbot, well. It. Can help with customer support because it can have less people trying. To man like phone, doing dumb calls if they're just trying to work out where their order is up to or something or other so, it really is a matter of I I, really believe that yeah. Should everybody be doing all of it well no, you. Don't need all of it you need to work at your business problems your brand has, got to flow through those experience, and if you're all about customer support then you should have a great customer support story if, you're all about cheap, prices and you just want people to find you quickly then you should be picking the technology, that helps you find quickly on search yeah.
Pick, The technology, that suits your, business yeah. But you can afford to do them all right. Right. Well the exception, of maybe like the you know the billion, dollar companies, and. Yeah. It's, it's. Not. Like they have infinite. IT budgets, it's, just it's just not the case if they might have bigger IT budgets, but you know you're. Not going to get the return out of all these investments so right doing. It if it's not actually giving. You value, at the end of the day you're just creating more maintenance, and, more thing yeah. Yeah. You know it's funny like in even in the context, of like the, Adobe. Acquisition, in the Microsoft, acquisition, a. Lot. Of times the sort, of argument. Is that. You. Know hey we're gonna have access to more funds that are gonna allow us to grow. Faster, and do more and stuff like that and of course to a certain extent that's true but, I. Think I was listening to a podcast it, might have been with the founders of stripe. Recently. Where they were talking about how, you. Know yeah people are trying to figure out how to take money and turn it into innovation. That's basically what the entire VC industry is but it's, complicated sometimes you, know money doesn't exactly get converted into innovation. As. Efficiently. As you'd like and sometimes there's something. About that the people or the team or the organization, building it that's, sort of unique or magical. And so, I. You. Know I don't know I'm not always convinced by that argument that you. Know now, it. Can be true at times but, I actually think the greater value is synergies, it's. Sort of like if, I've got some product contact, them now because we're part of the same team we can talk more openly more, deeply, we can do deeper integrations, we can end up with a better experience you. Know Adobe, got some really impressive marketing, tools and web. No. Delivery, no, it's a digital, experience tools, and yeah. Rady of tools and all that's really great, but. They don't have a fantastic commerce. Back-end and so. If there's you know Magento. Is going to be able to get into markets, that it couldn't get into before it'll. Be taken more seriously as, a result so, it's those sorts of synergies that I see is the real value do, I expect after, paying a whole bucket, load of money for Magento they will then dump, the same bucket load of money into development right. Don't. Really expect that I think don't be and. I don't think that's really where the value is I think the real value that comes about is getting, products, aligned better making them work better together, getting. Things into market, so there wasn't before. Dad. I think is the real value yeah. Thoughts, on the, Microsoft. Github. Acquisition. I've. Thought about that one very much seems. To be acquisitions, going on all around the place.
Fully. Understand, but I haven't really thought about. I'm. Sure it's just be because, github, was such a cash cow that that's, why Microsoft. With. Microsoft, over recent years, know. Some. People may agree or disagree they used to be these violent, no. Religious. Wars, but. You know I was. Looking at some hardware recently, like the the Microsoft's, surface. To do that great being screen and oh it's. A really nice machine yeah. Yeah well done in the machine I don't think that nice now. I. Want to put my VR kids into it or something. Yeah. I I have, a couple. Of friends that have moved from from Mac to Windows, and, I've I've been somewhat. Seriously, considering, my computers like five years old now and, I've been wanting. To get something new for a couple of years but I can't I can't. Justify, getting that touch bar thing. On the I just I'm not I don't I don't buy it that, touch bar thing on the new Mac so. I'm. Not thrilled about the touch bar cuz I want that escape key because you just need it for bi, yeah. All. Right are, you still vias VI your, your main your, main, text. Editor of choice but. It's just always there and, so when you know you ssh. Onto a machine hey. It's an. Editor it's always there you know I I started. Before there was gooeys, yeah. University. I started on punch cards just to give you an idea of how ancient I am or how backwards Australia, is one of the other. You. Know back then we're. Just getting you know x11. Screens, just as I was finishing up University, whoa that. Was cool yeah. But. So, I think, you just use what you used, to yeah. I actually, want, to get like. You, to give us sort of like maybe, a little bit of an overview of your your. Whole career like you've been done so many interesting things but if, we have time we'll get to that we got some questions starting, to come in here a question, from Joe Hart. How. Does the, PWA. Versus a MP play out which, is I know sort of one of the big topics. We wanted to to, hit here you, know so. I can give a bit of a. Yammer. On a little bit about amp first and then I'll talk about how they yeah oh yeah, that'd be perfect yeah, so a bit of background about amp and you know you, know Google's. Caught a bit of flack about it and yeah, maybe it has no. Yes. He's my pitch or my way of putting it so. I'm spending a reason women a time an app at the moment going through and working, out the right ways to use it for e-commerce that's. One of the things I'm actually, spending a fair bit of time on right now so, amp originally, came out that most people know from publishing. Industry, and it was started. Off being static pages and the, whole goal was hey in Google search how.
Do You make this. How. Do you get the experience even better and, one of the ways of getting it even better was okay, well I can bring up the description they click on it and okay we'll send the user off to a site okay. You. Can get that site to go faster and faster and faster and then you get to a point of well. It's mainly Network delay or you know how do I go any faster than that well. The answer is how do you get faster than that is okay. Let's, use, something. Called machine, learning or, a, I actually, don't know how they do it but let's predict which page the, person is going to click on let's. Start fetching that page in advance. So. When they click on it we've, already got it and it can just go slam, and split, it up in the screen right, well and. It's a way of getting faster. And faster web. Experiences, and it's something you can do in a PWA yourself like on your own commerce site predict what, you think the most likely button they're going to click next is and preload, some of those assets if, you don't oh well well if the reason use is just reading a document just. Use your network a little bit so you've got those options to do it also in your own pwace, but. How do you do it for a Google search, well. That's. Where it starts getting hard because. You could just download the files and not do the rendering but, the rendering takes time - and. How. Do you make sure that the actual origin doesn't see that somebody is downloading, the files because if you get the users web browser to download it it. Actually starts, to leak privacy. Information. For. Example I go, to a site IKEA. And I. Get their little tracking pixel and great and so they start, remembering something about me and then I do a search about. Chairs. It, could be you know a week later of those search about chairs and. If. The web browser while, I'm searching Google pre-loaded. That IKEA page it's. Gonna come from my web browser yeah Kier now knows that I did a search for something and I on my, search results was a chair. What. Starts leaking information, I didn't tell Google, that I want to tell like here about it and, so this is where things like the app cache come in what they do is they anonymize, take. A they help it to go faster, they optimize the pages they do a number of different things but.
The Thing you can't do without is the anonymous cash in because, IKEA. Does not know who looked at that chair and. So. The Google cash can then do the preload or the Google search can do the preload from the Google the app cache to, say we, will not tell the, site who you are that you're looking for something and will preserve your privacy in the process and still, get the page you do it faster, couldn't. They in theory just. Prefetch. Just. Do regular. Vanilla, prefetching. Without. Passing, along your. Any. And. Without leaking like just doing a anonymized. Request, IKEA to prefetch or does. That prefetch, happen client-side. Because. It prefetches, client-side, right. Quickly. Browser. And so it's your browser's would be doing the pre search so, by doing the prefetch to the app cache. All. Right got. A love-hate relationship. And you know. All. That was saying how, can you do a good job now, use. Current technologies, don't worry about the future just try and work out what the problems are come, up with a solution and then. The. Internal goal which they might not have communicated, very well with students say okay now we've worked out a way that works let's. Generalize. It, come up with no open standards, around it and so forth and that's what you already started to hear about like things like web packaging. The. Whole idea of saying okay well if I can get content. From the origin, and I can cache it but signed by the origins, thing then, a browser can bring it up even though it's fetched out of a cache and say this came from the origin, because, it's signed by the origin. And. So these are the new no standards. That are now evolving and no, app has been invaluable, to. Solving some of those problems like, it's, got. Real-world experience, too what are the real problems what are not the problems while. Delivering a solution today that works yeah. Is it gonna take over the web no no. It's Google trying to speed up the web learning, along the way. Yeah. It's working pretty well well we'll. Try and learn and do better and so forth but bit longer, to plan there are quite. A few things and a lot of lessons been learned about what's worked, and what's not worked to speed up a page now make sure your JavaScript, doesn't block the render pipeline. And. Hey. Inline a small amount of CSS, overall that works better yeah. Defer. Loading of images behind all these sorts of techniques. Doing. Projects like app helps Google work out what works well and what doesn't work well and then, it can go off and talk to all the browser's and you know if they agree and then great, know as. You get the new technologies, out you, don't have to use amp anymore app will still be supported, there's. No reason to get rid of it actually it's got a number of other benefits as well it's sort of new simplifies. The. Management of, your site you can actually just build up a creaky, site with sliders, and carousels without running a single line of JavaScript now, you don't have to learn JavaScript, anymore just use this component, library and it'll. Be a fast responsive. No, good. Site, yeah. Possible, build a better site than out yeah, if you targeting. One particular thing and you've got fantastic, JavaScript. Developers, yes you can build something faster than a vamp is trying to be generic store but. If you're just trying to knock something out quickly. And. Will stop you will, discourage. You from doing things that will slow down the page and. So it's sort of like that's, to, me the sorts of value proposition about our and, so, yeah I know. That as a user, like, I love it. Well. You know if I'll and if I'm searching for something and I land on a page that's, that's. That's amp, it's. Just the I'm looking like you, were talking about you know out started with publishing. And that sort of one of the key use cases I can't. Stand you, know all. You. Know media, web sites right they've got so many other so heavy there's so I love. I love getting that quick lean, page that comes out at you right to the content. USA. Today site, and, what they did. You remember that try it was I don't, know if I saw that she's something GDP are related, to us a tear and so you would say today I, don't, know the reason but they didn't get all the tracking stuff, sort of that with all the acceptance. So they they, turned it off on the European side and yeah. Oh okay, yeah. Okay. Wow. Guess. What it goes a bit faster, yeah. Who. Did that get and. I mean they. They have to they have to support, their business model, of course actually, how does some just. Amp support, advertising. At all. Is. Now working out like no. And. Multiple fronts so part of it is the analytics, and tracking and trying to work out well how, do you do tracking without slowing down the user experience is. Okay. Tracking. Is important, you got to have it and, there's. No work, going on right now is around trying to paralyze, the tracking so in order to show, the use of the content but do the tracking in parallel, and, can you do that rather than do all the tracking because what some of the sites do at the moment is you redirect the user to a tracking, site number one and oh it redirects, you to another tracking site to collect, something else and it redirects, you to the real content yeah, you.
Get A few bounces, before you even get to the page and then the air the page can be in really heavy and so forth and so the idea of the parallel tracking is then a no taking user straight to the page but. In parallel, start. Doing the requests off to the tracking site so you still collect the tracking information. But. Do, it at the don't, delay, the user experience in the process which. Is also it's. A key thing of app as well because you've got to have the tracking information but, again, when you preload, you. Don't want to trigger the tracking then you. Want to trigger the tracking when the user brings the screen actually, clicks a link and brings the page up on the screen so tracking was actually fundamental, to app as well because. The. DES merchant was, the site should know when somebody's viewed it and clicked on it not, when it was in a search result. And so that's again part of what app allows it to do and why it sort of tied down is to sort of say no no, because, it's app we know how the tracking works we know we can fetch that page without any risk and we, know that when we bring it up on the webpage then, all the tracking will fire and you'll, get all the right things so it's a bang all the rules to do good tracking, but. There are challenges know people are still solving some of the problems and, not trying to pretend there are zero problems here there are they're, still working through some of those problems, but. A lot of work has been put into that tracking and continues to be put into their tracking and yeah, some of it is cross browser support, know it's things like you know Safari. If. You go off and hit the tracking site my, understanding is you know you get a cookie coming back. If. Depending, upon the domain name Safari might say well sorry, I'm not going to accept that cookie that wasn't a user interaction. That was a bit, of JavaScript, in the background whereas, other browsers, accept, it and so. If. You can kind of use cookie for session tracking you've got a problem because when. You come then through to your real main site, hey, that AB page, it, didn't load from me it. Loaded from the cache so. Whenever it she saw any requests and. Put, the tracking data so how do I know that that previous bit of tracking data that came through was, the same user that. Is now on my site and so, amp has got concepts our client, ID and so forth in it but you got to get all these things to, fit together and, know. So. And there. Are answers, to all these things but, when you try them out you discover different problems in real life and hey that's, the way the world works and great work through so, you can either take the approach if don't release anything until it's a hundred percent perfect or, you, can get out there you try things you learn from it and get. Real-world experience. And. You just try and do the best job you can and then you iterate on it and you keep working and you keep working and you, find the partners who are, willing. To take that slight more leading-edge risk because. Well, they get benefit out of it you get lightning-fast page loads. Well. Hey guess what a whole lot of people have found that increases. Their conversion, rates increases, their engagement.
Rates And a whole lot of things but. You know any new technology there's risk and Google and, part of my role is getting. Out there talking to number these people I'm not the only one doing it obviously that, feedback bring it back to the core team so I go off and talk to the core team that's it hey I found, out about this one I had a hackathon yesterday, up in San Francisco with. Using. App for, landing, pages for Adwords, because. Hey if you get you ad, come up faster that's a good experience, it makes perfect sense here, so, what are some of the challenges and so people were trying to adopt it on to their current sides and, you know some. Of the problems technology. Problems it's not amp it was well, I've got this current site I'm not quite sure how to merge that technology, in or I've, got this current site and they want it to look exactly the same or we've got our branding, guidelines how, do you make the app toolkit, behave, like that so you just work through these different problems yeah. Yeah. It's, interesting it's. Kind of like there's, this dilemma of you. Know there's. All these ways that like. You said there's some fundamental things that independent, websites just can't do like such as network speed and, prefetching. And things like that, but. Even, among. The things that they can't that. They can, do you. Know I think about some of the things I run into in the magento space with people that are on older. Versions, and they've got you know slower, websites, and you you pull them up on a Google PageSpeed test and you go and you guys have some work to do and then. You. Know you tell them what it's gonna cost because, like you said you need apps and talented, front-end, engineers to speed some of these things up and, sometimes. It's, not economical. For them so there's this dilemma, of hey. The web should be independent. There shouldn't be this one, monolithic. Site that's sort of running hosting, all the content, and yet, at the same time some of these things it's like well. It's so much easier for them to just kind of fix it on their end like hey like we can host this we like, we, can just fix not that everything's fixed but like we. Can make this thing run fast and give you a set of tools that you can use that you're not going to shoot yourself in the foot so.
It's, An interesting dilemma. Of. The things that I spend. Gonna be spending more and more time with but, shortly. After I arrived there was a CMS. Summer. And so they're doing Google. Is out reaching to the platforms, as well they're sort of saying how can we help like. Partly. Engaging. Given, guidance you know that's. Why Google, and magenta took on PWI for example Google had out reached to Magento, and sort of said hey you, know we're. Coming across lots and lots of Magento sites out there did. You know there's better front-end technologies, around and, that's how the conversation started off and you know that they provided. A lot more information about, how you can go about doing it and, I. Don't. Know if you know rowing me it would but he's spending quite a bit of time with no James, and they they they sync up and exchange ideas and there's, it's a constant ongoing conversation. Between the two organizations. And, so. Google. Is both doing things and it. Needs to do some short-term things but it's also investing, into the platforms, and talking to the platforms. And know. The platform's got their own roadmaps as well and so no not all of them want, the help at that time they've got other priorities yeah fair enough yeah, but it no, I do, spend time talking to platforms, and I'll give these step, number one was to make sure I knew what I was talking about before starting, to do that of course that, but. They're going out and talking to the various platforms to make, them aware of the, technologies, help, them learn, how to adopt them, because. But, it is a slow process as, we'd be discovering, you know it, takes a while to move the platform because, the platform's got an ecosystem, around it and so. You've got to try and help the ecosystem move. Along as well and so you know one thing we'd love is no feedback from no, agencies. What would be the fastest, way to help you learn to speed up a site any. Current sites as well like it's not all about the future the, future is a multi-year, roadmap. It's are slow it's an ongoing slog, know, Google. Spent a lot of time with WordPress, they're. Working, on doing app support, built in as any. Plug into WordPress, where you just turn it on and your site becomes, amp enabled. You. Don't have to do much you install plug-in you, know it's, a great, experience will. The same technique work for Magento yeah maybe, maybe, not. Now. What is the right way is it a matter of like is it going to ever work for my No, nobody's, going to get the investment back into an n1 site so, here's a. Matter of teaching people how to use tools like lighthouse is, it a matter of writing guide, books to speed up you know here's, what you should do for me images, and. And. So Google's actually quite interested and if anybody, wants to feed me back that sort of information, I'd, love to get it it's sort of like, I've. Got my own opinions but there's nothing like getting it from the horse's mouth like, if. Google. Actually one of the things I was shocked I didn't, really realize Google. Actually has got a whole team, massive. Team of people that goes out and helps people externally, to get. Their sites to run faster, no, and and. Developer. Relations is actually not the primary team doing that they've got like a consulting, group and it's more in the sales side of things and probably, ties into Adwords a lot and so forth but they Google. Wants, the web to be a good and open and healthy place it. Doesn't want to suck it all in it's don't achieve Google's, goals actually yeah. It's. Not what Google wants to do it was just simply a way of getting something up and going right now today it works, once, you can get it so that it can trust other cases, and make sure their content and make sure the content obeys certain rules and it's not gonna block that slow, down the queue because it can't just prefetch, any old page one, you got the privacy problem and two. There's. The risk that that javascript, is doing a bad thing and then all slow down the search page well. That, sort of defeats the purpose and so it's, it's working through these issues and trying to work out and so there's concepts like no. Feature. Policies, that. You know rather. Than saying thou shalt do app to get an optimized, experience thou, shalt support, these three feature policies, you allow small inline CSS you, will not have JavaScript of blocks the render queue you come up with some of these sorts of policies and then any Content that obeys the rules can.
Get The same treatment instead, of saying app because. At the moment it's those rules that we want but, the only way that we've got today, that works at guarantees them is if you use out we know you're following the rules because we designed out to follow the rules and. Part, of it was but let's build app and work out what the rules are and great now we can work out what what, the rules we should be doing yeah. Right. Could. You could you give sort of like an overview, of where amp, is that for, e-commerce. At the moment like I I've definitely as, a user. Run into content, websites, content. Pages on em totally. Get that workflow I don't. Know that I've actually, seen like ass like a store, on amp, so what, is that what does that flow look like like what sort of the current state of where things are right, I think the Coptic oh let, me go check that out you. Might need to do it on a mobile device. Recently, but so. Carved. Calm is. A Magento, site. Which. Is actually both using app for the landing page so, if you do a search for a carved, case. There's a good chance that the ad that, comes up will be an app page and then. When you click on it and go through to the site it's a PWA oh. Nice. And, so. You know it's. Quite a nice site and it's not the only one around but that's no there are sites already doing this today and so. Do. Have people built the whole experience, of the site all being amp well I haven't. Heard so many of them it is definitely. Got. Some real benefits. And it's the one of the ones I'm sort of exploring, at the moment but, it the. The. Reality of the real world is frequently I got a site how. Do I get amp, added, to my site and so if, you're just adding out to the site that's where things like paired landing, pages to say okay, I'm going to create a page I, don't. Want to touch the whole site I don't wanna rebuild the whole site I'm just want to try this out I'll get a paired, landing page up and going I'm. Gonna make it look like my main site. But. I should be able to have, that in app until a lot of people are doing that and that gets you the first load, from Google search to your site to, be much faster and so they can click on it but then how do we get oh but, I want to pick the colors do. I have to click and go to another page before I can start interacting and so, what amp has added over time in its like a year, or two ago now it's not like new they've, got a lot more interactivity. You can now do things, like pick colors, and it will do a dynamic refresh. Of an API call and get back to data and refresh the page so it changes all the images, over. But. It's not full JavaScript they've, got their own. Language. Built into it to stop you from doing, anything it's going to risk the. Page and they're looking at JavaScript they're looking now to sort of say hey can we loosen up the restrictions, and still, give, people more flexibility, and so it's it's not like amp is a dead project that's. Static they're. Adding more and more and more components, to it all the time that more than this built-in. Carousels, just say amp, carousel throw, all the things in the carousel and great you've got a carousel that can do all the swiping, you stick a few attributes on to it and, you can control it you can do no. Checkboxes. And radio buttons and you can click, buttons and read load state from the server and dynamically, refresh, the page. That's. All there and so, you can do. A lot of it but, then there are still little, gaps like okay how. Do i display what my cart is and it's. Not that the technology, doesn't support it but, you need to glue it into your current site and so. If I'm going to do an add to cart button that's. Where it gets a little bit more work. There's. Lots of things part. Of things that I'm gonna be potentially, looking at is like okay, well, how do you do a landing, page for app and then make add to cart work nicely or is.
The Better solution change. Your product, page to, be at conformant. And. So there's some of the things that I'm exploring, for, Magento, and other platforms is trying to work out what, is the right story and you. Know and fiddling. Around with these tools to try and work out how, do you get this to work together nicely yeah right. Yeah. Architectural, how do you balance out those trade-offs of having full control over your product page, versus. Fitting into something. Like. Amp which probably gives you a lot of best practices out of the box but also gives. You a bit of a dependency, there yeah, and how do you do it in a cost-effective way that people can't afford to adopt because you come along and say oh yeah, amp is great redo your whole website. Yeah. Right I mean who can afford to do that just to get amp and the answer to me is if, most of your traffic is coming from app and it's a lot I'm, sorry from Google search and from other search engines it may actually be worth doing. But, if it's not and, may not be worth doing no like. There's any technology. That says oh everybody. Should just do this one regardless, it's. Just not the case right, you, can help you it, can give huge conversions you get some people getting spectacular. Results, well, everybody get the same results no. Right. So. I was playing with carved a little bit when I go, to carve, calm I see the nice, PWA. I'm trying to find the. Amp landing, pages like if I do a search like carved, case I see. Some search results for. For example a carved. IPhone case and it's a I think that's a Google Shopping and then link. And then that lands me at carve calm so I don't think I'm actually hitting any of the the. Amp pages we. Could I could figure this out offline, but it's. It's it's good to have a nice use case to to refer to yeah. Where, they actually went through how they actually built it all got. It I'll find that and link it up yeah.
So. Then. You. Know so that was yeah, so, that's, a great overview of kind of where amp. Is at overall it sounds like maybe. Most, of the time you're, maybe thinking more, about having your landing pages maybe. From ads or different, things like that land on your amp pages and then, patent. And because it's common for people to want to try it before committing, their whole site to it. It's. Got problems though and so this is one of the areas that personally. I'm sort of digging into with more is to try and work out the right story because you know one of the bits of feedback we got was oh great I've got two versions of my page now to maintain right. It's, just one and, so, it may be then that know what I'm starting to do is look at some of the e-commerce platforms, and saying hey, this platform you. Can convert your main page to app this, platform well. It's a bit of a pain so it's, probably it's probably better just to stay with a paired landing, page and then. Okay. Well, how do we educate, people and how to achieve it do we do some proof of concepts, do we build a Magento theme and. If, there's different strategies we could use and that some of the things we're thinking through at the moment so, it's not like this is a finished. Area yeah. But, lots of people are using our blots of people getting success, out of it. But. As. I said no technology is a solution, to everything, what's. Most important to you business is it is, getting more SEO like. It's not improving your ranking, it, delivers a faster, experience. We. Know that if faster. Experiences, the slower the experiences, people abandon. The page, and. So it's, not like do app because it'll boost your Google, rankings that's, not actually, the reason if it can be a a. Clear-cut, argument. To get hire people to listen but. It's more about you, know if, you don't respond in three seconds fifty percent of the people will just give up yeah. Which. At the end of the day is what Google's trying, to model is is, what, are what's. Most useful to, end-users and they've. Done a pretty darn good job of that over, over time, and. So yeah so if you just, focus on improving. Your end-user experience, exactly. That's we're probably gonna yeah. Okay. Well now we've got the technology how do you get it rolled out everywhere how do you educate people, on it and it's, it's. A problem, especially at scale and it's yeah and. So we know, Google looks at short-term solutions, and it looks at long-term solutions, it, does. Both right. Yeah. It's. Interesting you know you mentioned the idea of hey could, you throw your whole site on amp that, you'd, have some limitations, but that, would be kind, of an interesting approach, it's. Interesting because people, can, have entire businesses, that are just on social. Media, that are just on. I'm starting. To get a bunch of background noise over there oh. That's, better now can. You still, there. Yeah. Oh yeah it just got. Kind of fuzzy. Okay. Hopefully they'll clear, up but yeah. I mean you can have your whole, whole. Business on your. YouTube page and your Facebook. And Instagram and. You, can almost have entire, businesses, built on. Other. Platforms. And. So the idea of having your whole your whole e-commerce site on amp is it's kind of interesting actually probably, you. Know for some scrappy. Business. Out there that wanted to try that out play with it, that's. Good. Using. Speed. Up development no. Playing, around if you can view or. A completely server side at the time so what. Is. View. But. Allows you to build a series of pages that you can navigate the plane but use modern.
Rendering. Technology but, builder can put the server-side generated. Side so it's all server-side, rendered but still using. You. Components, perhaps that's. A liar error on top but anything can use UV technology to, build up the branding, and your component, library for the styling, symbol. A server-side, side, or. And. The. Reason for that the server side is. It. Helps with it search. But. The page. Pre-rendered. Okay, put a blog out with a couple of other days. Google. Can index drivers. To bring pages, and it's going to put more effort into doing a better and better job of it but it's add a little, bit of delay in the pipeline and if, you've got a really dynamic site like the laser paint if, you want to get rid of that delay the best solution, today is, to have a server-side, record and. So why. Not have the whole site, server-side, render than all air so. This is some of the ideas of sort of playing around with this. But. Is. That better than building and give the theme it's. All bad faith because. It's. A bit of work with what mobile people are. And. They. Spent a couple of days on it I think and built and then to no one, page got it up it wasn't very pretty but. It's not actually that hard what you do. Right. Right. Right. By. The way Andrew, in the chat did mention that. You. You appear to be fighting with the snow started. Here your, bike ice I thought it might have just been me but I think I don't, know if maybe you could unplug and replug you're, my king or something like that it got. That. Sounds better. See. Not. Hearing the snow anymore but. Now night not hearing you either. Still not even think, we'll. Get this figured out here. Still. Not hearing, anything. See. Still. Not hearing anything doesn't, seem like you're muted cuz usually if you're muted, it'll show a, mute. Icon. Thanks. Laura digging the recliner. The. Best. It'll. Cause it my burqa, lounger. Still. Not hearing, anything, it's. Really, strange. Um. Huh. It's, really weird I wonder. If you if you have no mic, at all if it might work okay, it looks like he's gonna reconnect. Just. Jumped out he's gonna reconnect. Kid. Revert yeah. Hey. Ah. I. Heard. Something but I'm, not, hearing you now as I, see you talking. Okay. Now you're muted, now it's showing me a mute icon, okay, now the mute icon is gone. Still. Can't hear anything. Oh, I. Think I could I think I heard something, Hey. Oh. Perfect. Okay, you sound great awesome. Rebooted. Different, headset hopefully this works better there, we go yeah that sounds actually amazing, so. Sorry about that so. Sorry. For my am. So. I can't remember where we were but. So, think about a amp. Amp. Is great how's. It to summarize things. It's. Good. It's. Got its users uses, and it's. Like, the the PWA. I think, we're started getting integration. Talking. About next, as. Well. The. Complete side so it's that's something I'm having a bit of a play with so yeah I dabble, again and part. Of is trying to work out, well. At the end of the day if, you, can come up with a really clever technological. Solution, that nobody can adopt, that's. Not, particularly useful there. So no I'm I'm, very. Much interested in getting feedback of. For. Those who want to implement amp what. Is the best way would, people find it more useful to have an app theme. For Magento one or magenta two would. People find it better to have a separate gateway that talks api's. Back-end. I'm. Very open very interested, to find, out more so thats some dansgame, it's. Like that's really current that's what I'm doing right now is just sort of dabbling.
With Some of these technologies to work at what's possible okay. Yeah, no we'll definitely be, pushing, this out there and try to get you, as. Much, as. Much feedback on that as we can you you, said one phrase I wasn't quite sure what it was you said a paired landing. Page so. One. Approach with, amp is to say create, a separate page and so. I don't touch my main site you, know I don't risk anything on my main site I'm gonna build, a new page with a different, URL it. Might have slash amp on the end or amp. As part of the domain name or. Something like that. And so, god be my new page I can build that up I can get it I can test it I can and great when it's when, I'm convinced, that's the right thing to do I can send all my traffic to it instead from for. Adwords, or what. However, you want to organic. Search and so forth and so, the, main page has a link, to the amp version, and the. Amp version has a link to, the the canonical, the main page. And so that's how the, two work together and that's how the crawler and so for, example you. Don't normally get mobile amp pages coming up it's. It's. More for mobile, accelerated. Mobile pages all, the pages and. You. Know it's. Just but there's all sorts of ramifications are, there and no long term should. We get on to the desktop hey interesting, question and no, do I have all the answers no I don't have all the answers I, don't, control it all I'm just part. Of a big, and large organization, working, through all these issues with teams of people behind it all yeah. What. Are the like. What, are some of the results and I know, you're not like in the marketing department but. What, are some of the results that like you mentioned carved, or other. Merchants, that are investing. In getting on some. Of their pages landing page lamp, Aliexpress. Switchy, remember, all the numbers I probably should bring up a web browser but, there's a whole case studies site but, I can't say on the internal mailing lists I frequently, see no. Site eggs got you know double D conversion rates so. Doubling. Is not shocking I. Suspect. It may depend upon how bad your site is at the moment, yeah. There's, always that fact you, might get this benefit from it but. You know some, sites have. Have. Difficulties, and challenges one. Of the things are discovered if you do a crappy, amp'd, landing page, you. Don't get good results. Just. In. Yeah. Absolutely, some. People sort of say oh wait up. A quick aunt page and and it doesn't look anything like this pretty is well designed as a main site and they say oh my results, were terrible, guess. What you, amp, is not gonna solve UX no, UX. Matters speed, matters they both matter. There's. Not nice formulas. To tell you oh you. Expen is 23%, point better, than speed. Don't. Have that level of granularity of, data at this stage unfortunately. But, speed. Definitely. We know people give up on sites if speed, is not there, no. Doubt about it. It's, got lots of metrics on that but also. Know that if you do a crappy looking site that. Just, on your main site UX. Matters, on your main site how you lay out your page matters, where the call-to-action buttons. Matters. Saw. One page and I looked at the two pages I thought gee, it really made that much of a difference and, sometimes. The changes are shocking. So, there was one site that had, quite a bit of detail on one image on the site and then, they changed it over to three images where. You clicked on one of the three images and it took you through to the area and no, engagement. Doubled and. It was just due to the way the page was laid, out and it was just clear at a workout at oh okay I've got these three choices click I'll go through the next page and I'll go into that next level it, was too scary having too much detail for a, sub-segment. Of the. Market and oh it, wasn't the image on top that got clipped the most it was the image at the bottom Oh. Guess. What and it's interesting thing and I've read a bit of research recently and I haven't really got my head completely around it you. Notice apps tend, to put the buttons at the top bottom. Yeah, don't, but. Mobile sites still, tend to put them at the top right. That, is interesting. Hmm. Well, there's a reason people put the buttons at the bottom they're. Easier to get to with your thumb, right. Right. But, we don't seem to have made that transition yet, on websites. Right. For. Mobile obviously. For a desktop, I'm, not saying buttons at the bottom of the right thing for a desktop. Those, lots of interesting things you know I remember all that discussion also with um you know the Dux team at Magento, we've, got the tabs down the side Hey, well it's actually quite convenient for a tablet, and.
It Works well on a desktop and so you've got this good, sort of mix between the two might. Not be the right thing for a mobile device well, guess what the, admin interface it probably wouldn't try and a mobile device is too much page. Not anyway. Do, you do you have full control over, or relatively, full control over like CSS, within. Within. An app there, are restrictions so, there are things you're not allowed to do in the CSS, but yes it, is HTML. It's a set, of library, of web. Components and. A. Set of restrictions to say and. It's really sorta like don't use your own JavaScript you, can mess it up too easily if you do the wrong JavaScript, or block the page render and it's all this cascading, sort of thing like oh I've. Downloaded the HTML, I've started parsing and said oh I need a JavaScript, library okay we'll go fetch that okay. I'll go fetch that I've got the library great oh we've, got to start executing, it because they've got to be the code on the page and I don't know whether it's safe to render yet I'll go run to that code okay go run that oh it needs some more JavaScript okay go go fetch another JavaScript and here's, cascading. Ripple and so, it's things like. No. Don't. Try try to avoid that try to load things asynchronously so. It doesn't block the main render, thread try and and so. I'm, not saying you can get rid of everything but it's really a b
2018-06-09 22:34
