Frontend Cloud: Unleashing Creative Power

Carlos Ponce (00:00):

Good morning everyone. Welcome to another episode of Tech Leaders Unplugged, and I am joined today by my fellow teammate, Bill Vannerus from LogiGear. Thank you Bill for, for graciously co-hosting today's conversation.

Bill Vannerus (00:26):

Yeah, glad to be here. Here, Carlos. Thank you.

Carlos Ponce (00:28):

Yes, thank you so much, Bill. And of course, we have our guest, Lindsey Simon, the VP of Engineering at Vercel. So Lindsey, thank you for joining us today.

Lindsey Simon (00:39):

Thanks for having me.

Carlos Ponce (00:41):

Absolutely. It, it'll be a, a blast, I'm sure. So, okay, first of all, let's get started with you, Lindsey. Tell us, first of all, tell us in the audience a little bit about you. I mean, we're, where you're coming from, your background, anything you want to say about yourself, and then tell us about Vercel then, and then we'll, then we'll move on to the topic. Thank you.

Lindsey Simon (01:03):

Alright. Yeah, great. So, my background, I got into programming for a couple of different reasons 25 years ago. But one of the things that happened, my first real job was as a webmaster I worked, and I don't think that's a job that title that really exists anymore, but at the time it was it felt pretty d and d cool to me. And I ran the DNS servers, the mail servers, the web servers for a newspaper, a weekly newspaper called the Austin Chronicle. And it was, you know, it's largely an arts and entertainment magazine, and I felt like my job was to provide a platform for writers and for creatives to, you know, share their insights and share things about other artists and creators with the world and politics, news, et cetera. It was wonderful. It was really coming from the point of view of being embedded in the community. And the web was pretty nascent at this point. You know, we were wondering if the web would ever be a thing that anyone ever made money on at this time. So that just kind of grounds the time and place, I think. But, you know, the way I kind of cut my teeth was by starting to experiment with Apache and Mod Pearl. And then ultimately PHP it was this huge revolution of being able to make dynamic real time web applications. And so, you know, the practical value there was, oh, okay, a writer wants to correct a story, you know, in a newspaper you have to wait till the next week and then issue corrections and on the web we could just change it right now. And, you know, using PHP read from the database and it was just live, you know, we would generate pages. We had static content generation even at that time, and it was very powerful. So that really set up the stage of, then ultimately, I went to go work at startups and got to Google about 15 years ago and started working on a pro, a tool there called App Engine. And Appen Engine was very similar in its time and place, and still is to the mission of Vercel. So, this is kind of a, you know, a long-term dot connector. But App Engine's purpose was, we're going to give you a framework and at that time see Java and Python, and you can build an application on your machine with any resources you want to, you can debug it, you can reason about it. And then you type a command in your terminal. Magic happens, you know, in this case it was like an SYC up to a server, and then you get a URL hosted on Google infrastructure, you know, best in class CDN, best in class routing. You know, was a very comprehensive set of tools. And I think that was both its benefit in its reasons for some of its demise. It was kind of a locked in infrastructure. And then you fast forward, you know, almost 10 years and I had become pretty enamored with Next JS and what Next JS was trying to do because I saw the opportunity to build into a framework, all the things I tried to do in a build system to optimize websites, and it was just like baked into next JS. So I started experimenting and playing with Next JS really early on, and became a customer of the platform at that time called Zeit. And, you know, I don't know, a year later I was working as the VP of engineering at Quip owned by Salesforce, working on mostly, you know, infrastructure and, you know, really large customer needs, enterprise needs there. But I kept a side project going and I kept it on Vercel and I had this incredible customer success experience from someone who's very technical on that team. And, you know, I pushed a ticket in and said, all right, I can't figure out why my build isn't working. I've been paying forever. I know I don't technically get support, but maybe you can help me out. And one of the CSS people started pushing changes to a branch up to my public repository on GitHub, which would trigger a build. And then they would push another one to change, and that would trigger another build. And I thought this was the most incredible technical support experience I'd ever seen in my career. I wrote into jobs at Zeit.co and I, I wrote something like, Hey, you know, here's my background. I don't know if you, you know, what your plans are as a company, but I want to work there. Fast forward, I start working at Vercel. I got a job in product. Initially Ermo told me, you know, we, I just sight unseen, wanted to work here, and Ermo told me at two weeks before I'd start that he wanted me to work on product. And I, you know, I got chills up my arms because I've never been a product manager before and said, oh, yeah, great. That sounds amazing. And, you know, got off the call, freaked out and started figuring out what I'd do next. So I started talking to customers right away when I joined the company and getting to know people, but talking to customers was, I think the thing that really underlies all of what's made Vercel and next JS successful over this time. It is a true dedication to helping people succeed on the web with a combination of open source and services and infrastructure.

Bill Vannerus (06:19):

Picking up on that's a real key point Lindsey, and I know on the delivery side where, where I work you know, listening to customers' needs and, you know, the issues and challenges that they face and addressing those is key. And I can see how excited you are about, you know, the firsthand experiences. And you, you mentioned a couple use cases there where solutions problems were, were festering and solutions were found, and the excitement that you found in a key solution that that really drove to, you know, improved efficiency and productivity for, for that development team. So could you, could you talk a little bit about, I know in the, in the software industry that we're in all the way from, you know, business influencers decision makers, all the way down to product managers, product project managers, even Scrum masters, everybody's looking to really increase efficiency productivity while still providing creativity for development teams and individual developers. Could you share a little bit how the solutions that, that you and your team provide really help the next generation of, you know, improvement in efficiency and, and increased, you know, well, not stifling creativity, right. Taking things to the next level?

Lindsey Simon (07:46):

Yeah. That is the, I think that's the, that's the thing that connects most of what Vercel is trying to accomplish. Yeah. You know, we have an internal metric, which we call our confetti metric. It's the time to happiness. Basically, it's the delta between you landing code and us giving you a U R L. And we try very hard to focus on that performance, but that performance is part of the whole, the premise that there are a set of tools, there are a set of capabilities and, you know, some norms around what it means to develop successfully for the web, which we have sort of collectively now started calling the front-end cloud. And this is a, you know, kind of a segment off, okay, get compute and things into the cloud. It really started with get VMs, get machines into the cloud, and now it's, do you know, something like a cloud development lifestyle, right? Like, try to take advantage of the performance you can get on the network to build better together. And that is what, you know, it's a, it is a focus on accessibility, design, speed, usability, these, these are all things your developers want to be able to do, and to get the recipe of things together so that you can successfully and consistently deliver an amazing user experience. Those are the tools that we've bundled into the terminology of frontend cloud. Yeah, you know, I think a lot of Vercel comes from a lot of people who work here and come from a background of working with either agencies or doing work for friends and family. And, and this is where it all kind of comes together into the business, right? The business ROI is, you don't need to spend an arm and a leg on custom bespoke infrastructure, which takes time, money, and, you know, a lot of different kinds of resources to do something that, you know, really large companies have figured out in some ways. And, this has been codified, and we are giving this now to hobby developers and small businesses, and large enterprises.

Bill Vannerus (09:55):

Okay. That was kind of my next area I was going to go into. You know, what, what are the different types, of companies that come. Is it, you know, large enterprise businesses, small tech startups, you know, blockchain web three? Cause I know we see in our enterprise customers, even some that are looking to, to again, try to retool where they may have, you know, tried to make the shift from, you know, large client, server-based legacy suites of apps to, you know, cloud-hosted, quote unquote you know, putting a web interface or trying to even migrate some of their key apps to SaaS and, and trying to pivot to, you know, software as a service versus just releasing, you know, large legacy enterprise applications. So, what do you see a broad spectrum of companies coming to the table and what are the, what are the type of challenges that they face specifically? Are they the same types of challenges between large enterprise, small tech startups, hobby, you know, developers, like you say, is it similar or is there a lot of differences?

Lindsey Simon (11:11):

I think there's, there are a collection of similarities and, and then of course there are differences, right? If you're a hobby developer or a really small business, you get to start, you, you have the, you know, your greenfield, you get to pick the latest and greatest tech stack of today and, you know, go from there. And if you're coming in and you've been in business for 10 years, or you're a large agency, right? You have customers on all different kinds of tech stacks for people starting from scratch. I mean, there's a very healthy flow of people who start for free on Vercel, decide to build a startup or something like that, and then that business gets traction. And that kind of comes back to the idea that Vercel is designed to make it really easy for you to push up a high-performing, globally distributed website, and business from day one. You know, you don't have to go refactor and production. If you adopt this kind of stack today, you're there. And where enterprises come in is very much from the point of view that's, you know, been go, you know, the term digital transformation has been going on. You know, this is a while now, and I think this comes back to a couple of really key things for enterprises. They have a developer team who knows what the latest and greatest is, because it's very important for developers to stay on top of trends like this. And then they come to work and they're working in a system where, you know, C I C D is weekly, there's a build cop rotation. You know, they're, they're just disappointed with that. And, and so, you know, engineering leaders both want to change for that reason, and they want to iterate faster, right? All evidence points to being able to iterate faster is the key ingredient to success, right? And so, you know, so that means enterprises come to us and they often say, Hey, we want to get there. We have one part of our site. Maybe it's their docs, maybe it's their app, maybe it's, you know, some set of marketing sites and they want to start making this transition. So next JSS design and I think this was one of GMO's, you know, very early position for next was that you could do this on a page-by-page basis. You could iterate, and it wouldn't be like, oh, you need a gigantic JSS bundle for this page now. Or the same thing you had on other pages. This could be lightweight, allowing you to say, okay, move part of your site over easily. That, that was a, that's a big deal because then you've got, you know, this beautiful internal competition, right? You've got a team who goes, hey, look how much happier or better we can do on this new setup? And they really validate it. We don't have to prove it to them or, you know, we can tell them it's better, but, you know, if you can start, there's a comment around build previews. If you can start to iterate more quickly, that is everything, right? That time to live is the endorphin rush for a software developer. You know, if there's nothing worse than leaving a bug online for a week that you know your friends are, you know, customers are hitting because you have to wait to push a build till it passes, you know, a million QA steps and yeah. Etcetera. So that internal competition leads teams to start to push towards the, the transformation of their tech stacks and their tools.

Bill Vannerus (14:33):

Yeah, that, that's a great point. I know I was reading on your website where you guys provide you know, in your framework, in your platform, the functionality to really have that preview prototype of, you know, before you merge your code into the, you know, master. Could you talk a little bit more about how that improved efficiency? Do you guys track any kind of metrics or KPIs that really show, you know, the improvement in, in efficiency and finding bugs sooner in the software development lifecycle or the path to production pipeline?

Lindsey Simon (15:15):

Yeah, I mean, you, the easiest example of this is in the code review process. You know, the prior state of the art was I send you a, you know, a branch and a pr and you to test it now need to pull down that branch and PR stop, whatever you're doing, go pull that branch and PR down and start testing on your local machine. And the, you know, the number of varieties things that can go wrong when you do that is, you know, many like, usually, and, and so what happens is people skip that step, right? Then you just look at the code and try to think, okay, does it work? Does it look like it works? I don't think that's very effective. You know, instead you get a poll request and a link that you can immediately go to and test in real time whether something works and have a actually very reasoned conversation. One of the reasons I came to Vercel was to build out a feature that we've now had for seven months, eight months, called preview comments. And the idea is on these previews, which are generated automatically from branches, right? On these previews, you can go and leave comments and like, you know, say, hey, this, this looks weird on my machine. And in the past, right? You, if you knew how you'd take a screenshot and send an email, and then you'd have to follow up with, well, what browser were you on? What did you know, an, an infinity of questions that all get back to you. Oh, you were on the wrong URL. And if you go to a stable URL, like a branch name, and you're always seeing the latest and greatest, and you leave a comment directly there, we capture all that metadata, and now you have a real time bi-directional syncing conversation that can post back into Slack or other tools you use, and it just changes the feedback loop. It speeds it up so dramatically. And so that both empowers engineer to engineer collaboration, but it also opens up collaboration to people in product people, customers, right? Like if you're an agency that that part of the process can be so painful, and you can make it so much better. And other personas who are previously sort of unempowered to engage in the development process, like maybe they don't have a GitHub account if you're a marketing exec, right? But you surely want to review the thing that's going to go live tomorrow. You can both leave meaningful input where it matters and see the velocity of it being dealt with in code by your team. Yeah, that's, and that gives you trust and confidence in your team and in the platform itself. So, it's very powerful.

Bill Vannerus (17:43):

Yeah. That's pretty awesome. I, I think I read in your site how you guys have integrated Slack into that collaboration loop, right? And I guess that that allows, you know, other non-tech technical people to get involved in in the dialogue. It's, it sounds like it's almost like a pre identification of, of defects and do you see in project teams that you can see a marked reduction in defects?

Lindsey Simon (18:15):

Yeah, certainly. I mean, it, I think the thing is you actually act on them. We all see the defects. It's when do we see them and how quickly do they get acted on? And I think we're pushing for see them first before they're live in front of your customers and the ability to act on them much more quickly. If you tighten that feedback loop, you're certainly going to make it easier to ship better products.

Bill Vannerus (18:41):

Yeah, absolutely. In that preview, do you allow, can you run your unit tests or integration tests, certainly along with.

Lindsey Simon (18:50):

A lot of people run a whole, a whole host of, you know, those previews are effectively, you know, because it's a branch, depending on how you cut staging or production or various stages of staging, using branches for that works pretty well. You can set up different kinds of environments on Vercel for your environment variables to be able to, you know, hit different data sources for different kinds of branches or stages in your development lifecycle. And it also doesn't enforce one way of doing work. Companies have established or, or not, or they haven't established, but, but a lot of them have established workflows that they want, you know, they're not ready to change that, but they do want to change the tools and eventually it turns out they change their processes along the way. And so we need to stay flexible with that.

Bill Vannerus (19:36):

Okay. Yeah, that's awesome. I know, I think you, you mentioned performance and I was going to ask you about, you know, kind of edge computing and, you know, regional deployment and the impact in, in that same kind of prototyping preview environment. Can you switch where your, your data's come from? Yeah. Or switch your, you know, where your, your URLs being hosted so you can test, you know, kind of regional performance for different clients, regional clients.

Lindsey Simon (20:04):

Absolutely. Well, I mean, one of the beauties of the Versal platform is that fundamentally, you know, we, we host your content globally, and then you might have, you know, you'll have, you may have compute in different regions, right? To get close to your database and you can determine where you, you know, send that compute. And I, ideally you want to optimize putting your compute as close as possible to your data stores. So that's totally configurable and configurable via environment variables in your previews. So, you know if you want to test different regions, that's completely doable.

Bill Vannerus (20:39):

Yeah, that's awesome.

Lindsey Simon (20:41):

And that, that enables your tools as well. Like if you have a tool that hits URLs in different regions and tests for Chrome web vitals, for example, you can validate that this is workings or, or tune and adjust. That's, and we report on that. We give you built-in analytics that tell you what your Chrome web vitals are in different regions, on different machines and, and from your users. This is another principle around web performance and testing that I've been enamored with for, you know, since my time at Google on the web performance team, being able to validate, you know, from real users what their performance is like. I I've spent a lot of time in my life trying to give people tools to do that. Cause that's how you make them better.

Bill Vannerus (21:25):

Yeah, absolutely. So, so your platform also provides the analytics real time in, in production environments and feeds that back and something like a product like New Relic might do?

Lindsey Simon (21:38):

We have a really interesting advantage over third party analytics though, because we are your service provider. When you, you know, if you load a script tag for New Relic or you know, any third-party analytics tool, immediately now you've impacted the loading process in the browser to like, okay, go fetch on a, you know, async or not hopefully async, but to go fetch this bundle of JavaScript now and then start sending analytics to this other server. Well, okay, now you've got third party subprocess right there, and you also have right ad blockers, right? Like you're losing intelligence right away, some percentage and, and a lot of platforms, this is a good thing, but if you're trying to make your product better, you actually need reliable analytics. And because we're your service provider, we ship an endpoint in your domain. So, if you are, you know, Vercel.com, we ship an endpoint underscore magic string, and that's where your analytics go. They go into first party ingestion pipeline that then we give you analytics, you know, a view of your analytics on, so you don't lose data in that process. And you know, you know, we make sure you don't send, you know, things you shouldn't in that as well, but you could send custom events and you can send, and we gather telemetry there that's reliable.

Bill Vannerus (22:53):

Wow. Yeah. That sounds really powerful. And we, if, if you were going to leave one thing for our audience and for potential, you know, customers, one thing that is really a key component to improving, you know, efficiency and performance of, you know, front end web development overall at what would that be?

Lindsey Simon (23:18):

Look at your data. I mean, this has always been the case, but we, we measure ourselves by our ability to deliver time to first bite for you as a customer of Vercel. And we look at that data as a measure of whether we're succeeding or not. And there's a lot of ways that we or you could, could, like, harm that user experience. But in the end of the day, you know, users are not here for the tools you choose. They're here for the outcomes. Our goal is to make it so that when users see URLs they don't cringe before they tap that URL, right? Like, you've been on clunky websites where you know, you know, you're going to just wait for it to load. And it's, it's, it's painful. And we're trying to deliver a different web and we want to bring everyone along with us.

Bill Vannerus (24:08):

Well, thanks. That's an awesome goal. And it's a really intelligent and well thought out and well experienced input for the community. I think anything for a particular customer that might be you know, still shifting to the cloud, they may be in still a hybrid mode where they're also looking at the, just the cost of hosting on the web, you know AWS or Azure, and they're really looking at cost reduction in those areas. You know, how can they streamline, you know, how can they be more efficient?

Lindsey Simon (24:47):

Yeah, I mean, that is the, that's the premise of why, again, what the thing that we believe makes Vercel unique. You, you know, the primitives for hosting or serving traffic you know, if you've ever spent time in the AWS console, you know, it's, it's extremely powerful and there's a lot of sharp edges. You can very, a lot of power comes great responsibility. And it isn't, I, I don't think most people pushing websites and wanting to create and develop, really want the added risk or the added DevOps costs associated with running infrastructure. And, and we see that that's why people come to Vercel when you look at the total cost of ownership for both the service providing, being provided, you know, down to something like bandwidth, right? But if you look at, okay, right, bandwidth plus I have to employ you know, DevOps engineers with experience in K8, I need SREs. You know, you all come, it all comes back to there's a better way. And we are that team for you.

Bill Vannerus (25:48):

Awesome. Well, it's been very enlightening speaking with you and listening to you, your insights Lindsey and sure appreciate your time here with us, and I think we're coming to the, the top of the hour here, right?

Carlos Ponce (26:01):

Yes, we're coming up on time, unfortunately. I mean, this was definitely very enlightening for those of us who, like me are more at the, at the layman's level with technology. But I want to thank you for having been with us today. I just have a quick announcement to make really quick. And this is about the next week's show. We're going to be, I'm sorry, on the 14th because we're going to have a two week a one week break for from the shows. We're going to be continuing on the 14th, which is a Monday. We're going to be speaking to Siddharth Shah, I hope I pronounced that correctly. The Vice President of Engineering at Canoo. And the topic is going to be software defined transportation, revolutionizing the road through software definition. So, stay tuned folks for another episode on the 14th of Tech Leaders Unplugged, right here at 9:30 Pacific. And join us here and stay tuned for more great upcoming shows on our website, techleadersunplugged.com. Thank you so much. Thank you, Lindsey. Thank you, Bill, and see you next time.