Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Impressive Hosting

Impressive Hosting artwork

How Can Performance Excellence Drive Managed Hosting Success?

Jul 1, 202536 minAudio

Jesse Friedman interviews Christos Paloukas from Pressidium about WordPress performance optimization and load testing. They discuss mobile performance challenges, Core Web Vitals, and how slow sites directly impact sales and SEO rankings. Christos explains testing methodologies, customer expectations, and why two-thirds of global users lack access to high-speed networks. Click here to watch a video of this episode. (00:00) - Teaser (00:24) - Introduction (05:39) - How does Pressidium do performance measurements? (10:03) - How do you set expectations for customers in load testing? (14:47) - What's the best way to help customers optimize for performance? (29:22) - What is the impact of performance on SEO in hosting? (34:52) - Conclusion

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show.

Teaser Christos Paloukas: Testing is really fun. It's basically you're DOS-ing things. Jesse Friedman: You have three seconds to convince somebody to do something with your website or they're gonna bail. Christos Paloukas: I don't really have a number for you, but maybe I'll confidently say two-thirds don't have access to 4G and 5G networks even today. Jesse Friedman: The second that page becomes dynamic, it's harder to cache, and that's where the real dangers start to come into place.

Introduction Jesse Friedman: Welcome to Impressive Hosting, where we seek to uncover the core tenets of great WordPress hosting. I am Jesse Friedman, and with me today is Christos Paloukas from Pressidium. We wanted to have him on the show today because he's working in the WordPress managed hosting space and has a lot of tips for us today. And also, we're all handing things back and forth here because I was on Christos's podcast and we thought it would be great to continue the conversation here as well. So, Christos, thank you so much for being here. How are you doing? Christos Paloukas: Good. Good. Thanks for having me, Jesse. We had an amazing talk on my podcast too, and I hope this one goes well too. Jesse Friedman: I'm sure it will. Christos Paloukas: Yeah, me too. Jesse Friedman: Yeah. So tell me, tell us a little bit about yourself. Give us the background of who you are, where you are, what you're working on. Christos Paloukas: Yeah. I'm Christos. I am based in Greece now. I was born and raised in the States outside of Chicago, and I came to Greece on April 22nd, 2010. I started working in IT professionally since 2012. I've always been a geek. I had access to a computer from my uncle in the States. He studied computer sciences, never did anything with it. And my first operating system was Windows 98 or Windows Millennium, I think. I don't remember. I had that amazing pinball game. Jesse Friedman: I love the pinball game. Christos Paloukas: That's, I think that's 98. Yeah, Jesse Friedman: Awesome game. Christos Paloukas: Who knows? Jesse Friedman: Yeah. So in IT I started off with systems as a bit of a systems administrator of sorts for Linux systems. I got involved with Linux 15 years ago. I was just talking about it earlier with this toxic relationship I have with Arch Linux on my desktop. I mean, it was great at first. I loved it, you know, playing around with window managers and making things fast, snappy tiling, and now I just want things to work and have buttons. Jesse Friedman: Decisions, not options. Is that the thing? Christos Paloukas: Pretty much, I mean, it's good, it's flexible, I don't know, silly rant. Some driver from my virtual camera to make all this fun green screen stuff happen got updated, but OBS didn't get updated, so I didn't have a camera and I had a podcast scheduled the same day, and I was dumb. I didn't Google. I could have found the recent bug report from the end of March, but anyway, I didn't, and I spent half a day troubleshooting because I chose Arch Linux bleeding edge. Anyway, systems administration. Yeah. I worked at a company called Codex. We did payment gateway integrations. That went well. It was probably one of my stays at work that turned into freelancing at a certain point. That office closed, freelancing lasted long enough for me to find a steady client that I ended up being their sys admin full-time, mobilized. And then that turned into an engineering manager position because I don't know. I love the team and I had this knack of being, what's the word? I don't know. Proactively aware of the situation when it comes to, okay, if the client wants a MEAN site, we're gonna give him a MEAN site. Even if it's a five-page website and they want to have MongoDB on the backend, fine. I don't care. You want Mongo? We'll give you Mongo. It was kind of hard to convince the developers at some points to be like, okay, we could do this in Laravel, but no, just use a MySQL database. But no, the client wanted to have MEAN written on the website anyway. Communications was always my, I don't know, strong point, actually humanizing IT, especially among the team. Jesse Friedman: I like that, Christos Paloukas: Yeah. Jesse Friedman: Humanizing IT, making it accessible in a sense, helping people understand it. Christos Paloukas: Yeah. Isn't it supposed to be like that? I mean it before, it was ICT, right? Information Communication Technology. Communication. But we kinda lost track. Jesse Friedman: You are right. Christos Paloukas: Yeah. It's basically Information CAT Technologies today, you know, the internet is for cats. Jesse Friedman: You said you started in freelance. I think that's a great thing about WordPress is that so many people have started in freelance taking on customers. They've built businesses up from understanding customer needs, which is a fantastic way to get going in the industry. Christos Paloukas: Definitely. Jesse Friedman: WordPress itself makes it accessible, Christos Paloukas: Really. Jesse Friedman: You know, a company doing something around it. Christos Paloukas: Yeah, it's difficult. It's difficult because you're a one-man show, you're sales, you're marketing, you are the guy talking to the client. You're support and you're the actual developer, admin, whatever. WordPress. So yeah, we should probably talk about WordPress. We both do managed WordPress hosting. WordPress. I would never took WordPress very seriously, to be honest, until I started working at Pressidium, and that's about just a little over three years ago now. Yeah. It was my go-to for websites because that's the way I learned in high school from Richard Nichols, my computer slash electronics teacher in high school. So yeah, it was always my go-to, but I started taking WordPress seriously from my performance engineer position at Pressidium. So what do I do at Pressidium? How does Pressidium do performance measurements? Christos Paloukas: I do performance audits on websites, whether it's front-end performance or backend performance load testing. Load testing is fun. I use JMeter at first. Now I switch to K6, because JMeter is great, but it's old. It's very old. I don't know if you have any experience with JMeter, but even trying to get HTTP/2 requests out on it, it's not fun. Anyway, K6 is great. You write code and it works. JMeter is kind of a, it's got a GUI, it's got one of those GUIs that's like some of the stuff that I used to see when I was growing up on Windows 98. But yeah, K6 is way more practical. There's been a bit of a shift over the years. Yeah, I can see years now over the years at Pressidium where I started working with the marketing team too. And yeah, now I kind of do a, I'm not gonna say 50-50 split, I'll say it anyway, a 50-50 split between the marketing team and the performance engineering for Pressidium. Most of the backend stuff is actually covered and has been covered by the SRE team and the DevOps team. So I'm there for the testing stuff. Jesse Friedman: So you are spending your time helping to make sure that Pressidium is running as fast as possible, that individual sites are going and optimized, but you're also working with the marketing team. Does that mean that you are trying to translate the technical operations that you're working on into marketable messages for end customers? What is it that you're doing with marketing? Christos Paloukas: That happens too. So, full disclosure at, WordCamp Athens? No, not at Athens, at WordCamp Europe. That happened in Athens in '24. I just had this natural knack of talking to people like, you know what we're doing right now. Jesse Friedman: Sure. Christos Paloukas: They were like, cool, let's do more of that. Do you want to, are you okay with that? And I was like, yeah, it sounds fun and it's still happening today. Jesse Friedman: So you're running the podcast, is that considered marketing? Christos Paloukas: I guess. Is it? The marketing team definitely considers it marketing. I consider it just talking to people. Jesse Friedman: That's right. Christos Paloukas: Yeah. Jesse Friedman: So what other aspects of the marketing are you touching? Christos Paloukas: Aside from being in the marketing closet with the people, I do communications. I talk on post data. I'm not in charge of things like sales or partnerships. We have dedicated people for that. Again, I may be a point of contact for someone or I may pass someone off. There's the, yeah, this is getting really full disclosure, but that's Pressidium's way. Plus there's like the personal Twitter that I started, which I wasn't on Twitter. I was, but I never used it. And you know, I just started using Twitter more and it's not so bad. I think I want to play around with like Bluesky. I made an account. I didn't go any further than making an account and Mastodon, but yeah, I just get the word out. Jesse Friedman: Bluesky. I feel like it's a green field of good conversations right now, but limited. We need more people engaged there. What are you talking about on Twitter? What's the handle and what is the general topic or thesis of what you're discussing on Twitter? Christos Paloukas: WordPress, myself, performance. Maybe I'll find something cool to retweet. Yeah. Do I do some Pressidium retweeting too? Naturally. Jesse Friedman: Right. Christos Paloukas: I wanna stick around like the whole performance side of WordPress without getting into the development side of things because I'm not a developer. Sure. I could write a couple lines of code. I could write some PHP, I could write scripts, maybe I could write small plugins, but no, no, no. I wouldn't consider myself a WordPress developer. Jesse Friedman: Gotcha. And what's the handle? Christos Paloukas: @paloukas. My last name, P-A-L-O-U-K-A-S. Jesse Friedman: Nice. Christos Paloukas: Yeah. Jesse Friedman: When you're working on performance, are you working on it from infrastructural network layer or are you working with individual customers and their specific sites? A little bit of both. How are you approaching performance? Christos Paloukas: Generally like in the real world for customers, particular customers, or not even necessarily customers. They might just be leads or they might just be people interested in like the free performance offer thing we've got going on. Jesse Friedman: Okay. Christos Paloukas: Which is cool because I get more experience with every website that I get to optimize because every real website is completely different. Completely different, and there's always fun surprises from the systems. Jesse Friedman: And helping to optimize individual customer websites, How do you set expectations for customers in load testing? Christos Paloukas: Yeah. There's also the system side of things, but the system side of things is more testing because again, I started at Pressidium like three years ago, three years and three months now. And most of the stuff was already there. Some hasn't been tested, especially for like, you know, large-scale load testing and that's where I popped in, like, hey, yeah, we could run some test routines and see where we can identify bottlenecks like all the traffic that we sent over. That was an experience too. I never had like, formal experience with load testing. Load testing is really fun. It's basically you're DOS-ing, you're DOS-ing things. Jesse Friedman: Yeah, we run into that all the time. When we work with, you know, obviously WP Cloud works with many different hosting providers, and Christos Paloukas: Yeah, first thing we wanna do is get them testing it because, you know, you can talk about it all you want, but the proof is in the pudding. And so Jesse Friedman: Yeah, have at it, but we have so many security layers in place to prevent DOS-ing and other attacks, you know, we're just like, give us a heads up so that we don't flag your load testing as something negative. Christos Paloukas: That's happened. Jesse Friedman: It's basically what you're doing, right? Like you're attacking your own infrastructure to see how it's gonna perform. Christos Paloukas: Yep. Jesse Friedman: How does this work at Pressidium? Like, does a customer sign off for a specific feature? Do they pay for something to get an analysis? Or is it bundled? How does it work? How do they get in touch with you? Christos Paloukas: We can actually, so like, what happens usually. Our DevOps support team, they'll get alerts if a site is going too slow or too slow for the particular site in question, and then they'll talk to the customer and be like, Hey, your site's slow. Do you wanna do something about it? And that's where I pop in. Or the other way around. Sure. There's sometimes that, you know, whatever automatic alerts don't actually work or maybe the customer's expectations are higher than what thresholds we've already set. And that's when the customer would engage with us and be like, Hey, my site's slow. Christos Paloukas: Which is probably the worst thing you could ever tell like anyone, any engineer of sorts. My site is slow because it doesn't really help me understand what is slow. Jesse Friedman: Right. I don't really know your expectations either. Like, okay, help me understand your problem, and yeah, we'll probably fix it, but it's a service that we offer for, well, the front-end performance stuff for anyone. The like load testing, it's for the premium series, the product line. There we go. There's the word, the premium product line, or the enterprise level. Christos Paloukas: The load testing because load testing does take time. There's like the study part, there's seeing what you, aside from expectations, what, where do your users go? Like, what sites are, what pages are visited. It's not optimal to, I don't know, do load testing on your homepage if you've got a bunch of ads all over the place and you're sending people to different parts of your website, so why would I load test your homepage if everyone is, I don't know, buying products on your WooCommerce site? Plus, that's where the fun stuff is, that all the uncached stuff, you know, when users actually log into your website. Jesse Friedman: Right, right. The second that page becomes dynamic, it's harder to cache, and that's where the real dangers start to come into place. So a customer says, they get an alert that says, or they're telling you my site's slow. It makes me think of that medical pain index screen where you see a bunch of emojis and the level of pain that they're feeling from one to 10, it's all relative, right? Because I might say my site is slow, and that's because it's loading in 500 milliseconds. But somebody might say it's slow and it's loading in three or four seconds. What's Christos Paloukas: Yeah. What's the best way to help customers optimize for performance? Jesse Friedman: What do you take to really diagnose what's going on with a site? What's something that someone could take home with them right now and think about? You know, the first step to really trying to apply some performance optimizations to their site. Christos Paloukas: Well, like any problem, I'd have to understand it first. So when it comes to like the, not so much the front-end stuff because, okay, let's do the front-end stuff first. The front-end stuff is pretty easy. I use Lighthouse or maybe PageSpeed Insights, but I try to stay away from PageSpeed Insights unless of course I'm trying to like send a link over to the client because PageSpeed Insights can give mixed results depending on, I don't know, unknown variables that have to do with the Google side of things. Whereas Lighthouse running on a local device or something like Sitespeed.io that I recently started using, we spun up a, it's a Docker image. Anyone can download it. Anyway, it also incorporates Lighthouse along with a bunch of other front-end oriented metrics, you know, like Core Web Vitals, all that fun stuff. When it comes to the front-end side of things, that's what I'll use. I'll scan whatever pages I should scan, including the homepage probably, but I'm definitely more focused on where the actual customers go to visit the website. Sure. The homepage might be important for ranking, but we still want to bring, or we still want to focus on the actual pages that the customers or customer, sorry, visitors are hitting. Jesse Friedman: Sure. Christos Paloukas: Then for the front-end stuff, Lighthouse does do some of the, well, I guess most of the initial heavy lifting. It's the suggestions and everything. Usually there's more behind the suggestions. Like when it says, okay, make critical CSS, that's not always possible, right? Plus, when we're talking about logged-in stuff, Lighthouse isn't necessarily the best tool. It's more of an experience-based thing for the back-end. Jesse Friedman: At home, Lighthouse works like right in the Chrome browser, right? You inspect your site, you go over to the far right, you see Lighthouse. It'll give you some information about the existence of the site, the way it's loaded, things that you can do to fix it. So you're talking about like critical CSS, like conjoining CSS files, pairing them down, serving them from a CDN or something like that. A lot of these times these things might be outside the expertise of a customer, so how do you balance, like what is the response? Let's lead this back into managed hosting. What do you feel is the responsibility of Pressidium to that end customer, to handle these things for them? Like, is it a matter of building a good foundation, and a good hosting layer for them that they can build on top of? Or are you holding their hand through this process to help them to provide better, a better experience? How far into this are you going? Christos Paloukas: I usually go further than I should. We have the same problem with the DevOps team. Most of the time they do some things like some deep-level code analysis to find where, I don't know, issues could arise. But where the demarcation point should be is me identifying the issues like, Hey, don't use a 4K image for your banner because it's gonna take long to load. And you're going to see, like if you do the Lighthouse thing at home, you know, if you push F12 on the browser and click on the tab that you mentioned, not the performance tab because that's gonna get confusing, the Lighthouse tab, what you mentioned. You're gonna see something like that. Something about improperly sized images. Or you're going to see like those really weird acronyms that are probably gonna scare you away at first, like LCP. Well, if you see LCP as in Largest Contentful Paint, that's the time when your browser finally makes it to rendering the largest image, not necessarily image, but it's usually an image, the largest part of what's visible above the fold, we say. So like the first screen they get, whether it's on the mobile or on their desktop, which is a good point. Typically the mobile versions are going to be significantly slower. Personally, I love doing optimization. Well, depending on the website, most websites follow a sort of pattern where they're similar, you know, the mobile version with the desktop version. Some websites might have a totally different mobile version when they're relatively similar. You know, like the menu changes or just like maybe the huge banner or some weird carousel things people still do. Maybe that's not there. I usually go off the mobile for the optimization because as you might see yourself, if you, for anyone watching. If you try running Lighthouse on a website, you'll probably see that your desktop score is going to be really high. The score is out of 100, so you're probably gonna see a 70, 80, 90 score. But on mobile, you might be surprised. Another really good combination is the way JavaScript files or like scripts and styles in general are handled. It's really easy to just defer all your JavaScript, and for most sites, that's fine. It usually works and it's going to save you a lot of time when it comes to rendering your page. So these words might sound really weird. Jesse Friedman: Deferring JavaScript for folks at home, right? Is waiting to load the JavaScript towards the end of the after first meaningful paint or whatever is that right? So that way you have something to see, some text to read before you load the JavaScript. Am I getting that right? Christos Paloukas: Yeah, just about. So if your page has something, JavaScript automates things on the front-end, right? So with the carousel example, for example, that I mentioned before. You don't necessarily need the carousel to be moving from the very first glimpse of the website, right? You just need the image to show. And you need the image to show fast. So you're gonna avoid like lazy loading because you don't want it to lazy load. In fact, you want to do the opposite. You want to set this little like fetch priority thing to make it high. So that way it loads before everything else. You could do even more, I guess you can, aside from deferring the JavaScript, you know, everything to render. And then the JavaScript might run is actually, this one's a dangerous one, executing the JavaScript files when someone interacts with the website. That's going to make your site. Jesse Friedman: You have to scroll or touch something or interact in some way. Christos Paloukas: Yeah, that probably, depending on your website and how complicated it is per page, because everything is per page, everything will change. If you go to a different page usually, that might break the site. That's the fun though, finding that like sweet spot, the middle ground. Jesse Friedman: You know, you touched on something that's always been stirring in my mind when I do a performance review of my site. You mentioned that you'll see a higher score for desktop and a lower score for mobile. Christos Paloukas: Usually. Jesse Friedman: Yeah. So like if I have both my desktop and my mobile phone on the same WiFi and the same internet connection, is that the variable that they're making an assumption that mobile devices are on some kind of carrier network that's gonna be naturally slower? Or is it that the processor of the phone is gonna take a little bit longer time to render things? Why is it that a mobile phone would be inherently slower than a desktop site? Christos Paloukas: Both of what you just said, so I don't remember what the default is like. Is it 1.44 or 3Mbps? Don't quote me on that for mobile. There, it's, what's the word? Not metered. What's the word for when you make something slower on purpose? Jesse Friedman: Oh. Yeah, it is I think it is metering or it sounds like maybe, I don't know, Christos Paloukas: Could be, anyway, throttling. There we go. So network throttling is something Lighthouse does do for the mobile and it also uses, like some mid-range phone CPU specs and RAM specs. RAM's really important too when it comes to loading a website. This is like nitty-gritty stuff that you could see in the performance tab on Chrome, for example, or Firefox. Let's not really get into that. It's complicated. But yeah, it has to do with both the network, definitely the network. And it's really important to note that like a huge part of the world, huge. I don't really have a number for you, but maybe I'll confidently say two-thirds don't have access to 4G and 5G networks even today. Jesse Friedman: Sure. Christos Paloukas: Or high-end cell phones. I mean most of the gaming market on the Play Store was targeted to the, well, was targeted. Not sure now to the Asia market and most of the people in like still developing or countries that have developed like not so long ago. They don't have consoles, gaming consoles. They didn't grow up with Nintendo Entertainment System back in the day to end up having like a GameCube if they're my age when they were going through school, and now they're playing games on the mobile, right? So mobile is a huge part, and the mobile is going to be slower on a global scale. But again, so this is another thing. What page your customers go to is important, but where your customers are actually accessing your site from is going to affect your actual real statistics. We talked about Lighthouse, right? So there's this thing called Core Web Vitals that runs behind this like Chrome User Experience program. When you use the browser Chrome on an Android phone, maybe, maybe an iOS device, maybe your actual desktop, and you give permission to have your data collected for performance analytics by the CrUX, CRUX it's called, kind of clever because it's got the UX user experience. Anyway, I like it. Project, what happens there is they collect data on certain metrics that Google tells us are important. It's a bit of a conversation there. Lots of techs might disagree because those metrics are heavily theoretical, sometimes like Interaction to Next Paint. It might not apply to actually like real-world scenarios, but it definitely gives you an indicator result or an indicator number of what your users are experiencing anyway. Once you stuff, all that stuff gets saved somewhere on the internet. And then it's used to rate like your website or webpage, that particular webpage. Right. All that information goes out. It's recorded now. You have a number, you have multiple numbers. You have multiple metrics for your website, and you could say, and you can actually see that your visitors are experiencing these numbers. If I run the test for my home, I might experience better numbers, worse numbers. What we wanna focus on is what your visitors are experiencing. Jesse Friedman: Yeah. So if you're an agency and you're working from your office, you may not, you may know to do this, but putting yourself in the shoes of your visitors is super important. We talk about this on this podcast all the time. Hosting companies need to put themselves in the shoes of their customers. Those agencies need to put themselves in the shoes of those visitors, so it doesn't help you to test your website on your mobile phone while you're in your office WiFi. You gotta disconnect and as you said, you gotta throttle, you gotta switch to 3G and see what it's like loading from that. You're not doing yourself any favors by thinking that your site's running screaming fast because you're on gigabit internet on your mobile device. Christos Paloukas: Total problem and misconception. Your customers gonna say it's slow and you're gonna say it's fast. But it might be slow for them. There's lots of tools online for like testing, things like testing different sizes, different screen sizes, how things will load. Definitely things agencies need to take into consideration when they're working with clients. Jesse Friedman: Yeah. So let me ask you this, and we're kind, we're, wow, this conversation's going so well, but we're already closing in on like the end of this episode. It's a great conversation but how do we, like, if I'm an agency and I'm building a website and I wanna make sure that the visitor is gonna have a good experience. Now with other things, a broken UX, for example, it usually gets surfaced pretty quickly. I can't click this link, or I keep logging in and it redirects me to something and it's broken. Like, you have three seconds to convince somebody to do something with your website or they're gonna bail, everybody's talking about that first little bit of time. So if people are just leaving, how is that information being surfaced up? And you know, a more apt analogy might be like if you were running a restaurant and your customers, you're not taking feedback from your customers and there's a dish that they don't like and they never return, it's hard for you to be able to pinpoint it as this dish was a bad meal. And so they never came back to my restaurant. How do you do that with a website? How do you connect with those visitors and also maybe look at it from analytical point of view or from a data point of view or something like that, where you're not actually interacting with them individually, but from a larger scale, how do you start to recognize that performance is playing an impact on the performance of the site, but the performance of your marketing or the performance of your sales or the effectiveness of your content? Christos Paloukas: That's a good one. So regardless of what you use for Google Analytics, if you use Google Analytics, you'll notice a high bounce rate. If you notice a high bounce rate from like users that have only visited your homepage or the page that you're campaigning or sending, you're basically screwed, like, sorry for the phrase, but it's not gonna work very well. It could happen there, it could happen when they're browsing in your shop. So if, say they've gotten to your store, they haven't logged in, they haven't had a reason to log in, and they're a new visitor. They were browsing on the products or they searched for something and then they bounced after the search. That's also a problem. Google tells us that if it takes longer than what, I don't remember, two to three seconds, they might bounce. If they bounce in the beginning, like marketing-wise, they're going to go to the next search entry on Google, and they're going to probably shop from the next store. You're going to lose sales from a slow site. Like it's not just a ranking thing. People really have to get that. Or it could happen at the very end too during the checkout process. So checkout is usually handled by some intermediary payment gateway like Stripe. You notice the order almost going through. But then payment isn't never made or isn't made, I don't know, for a good percentage. You know, something's wrong because you know, you already had that customer. They were, they were yours. They were just about to buy your product. But payment never went through. Why? There was never a chance. There was never an attempt. It's because they went to another website. It's because they closed the tab. What is the impact of performance on SEO in hosting? Jesse Friedman: And this is why performance is super important for SEO because when they bounce back to the Google search results, Google's keeping track of that. They're paying attention to the fact that you were only in that website for a few seconds and then you bailed, and now you went to another result. That subsequent result may now perform better than you. This is happening all the time. Christos Paloukas: Yeah. Jesse Friedman: That's why we try to tell end customers that performance is more than just providing that one good experience to a customer, but it's also about the effectiveness of your marketing, the SEO, everything else is tied to performance. Christos Paloukas: Definitely. Jesse Friedman: So Pressidium puts a major premium on performance, obviously. You're a dedicated employee who's working towards improving the performance of these sites. What do you run into in terms of guardrails? Are you enacting rules? Actually, you know what, before we jump into that, tell us a little bit about what your current customer, typical customers are like. Is it an experienced WordPress engineer, someone who's working at an agency? Or are you handling like DIY customers who are trying to do it themselves while they're running a store on their own? Where do you see your customer base? Christos Paloukas: There's definitely more small, naturally there's more smaller customers, right? So there's more like smaller, simple websites or smaller e-commerce shops. Lots of e-commerce shops too, you know, like people that, not necessarily on like the very small side of the entry plans, but like on the mid side of the entry plans, that's probably most of the customers are. Our focus would definitely not be so much on that end, but you know, aside from the money reasons, for the opportunity on the premium sites and on the enterprise sites. Like those are established customers or those are established brands? Those are established stores or LMS providers. That's a big one. LMS, anything that has to do with education. Jesse Friedman: Learning management systems? Christos Paloukas: Anything that has to do with education and people logging in and actually following along with the course. And what happens there is they'll come, they'll try us out, you know, they'll have that whole two week or maybe longer. We can extend a free trial period. And onboarding is like streamlined. Everything is taken care of by the engineering team for migrations, and that's some like Automattic tools because migrations can go really fast. But anyway. As part of the onboarding process, perform, like the other kind of performance testing that I mentioned before, the load testing is applied because when you have an e-commerce shop or you have a LMS website. You would need to, so, LMS websites have this issue. E-commerce shops are fine. They might have a burst of sales, a burst of customers, not during a sale period or during some campaign. But LMS users usually have classes, right? And if they have live classes, that means that you've got 30, 40, 50, 100 students logging in at the same time. You have to know that like the systems on the backend, this is where the backend plays a larger role than the front-end side of things. For testing, you need to know that the site's going to be able to handle those 100 students logging in, pushing buttons. Again, this is dynamic content. These are logged-in users. Caching is only gonna help you if you have like a very well-built site and you've applied things like fragment caching, that's when you cache certain parts of the page that you know are gonna be the same. But even then, at the end of the day, it's only gonna help you load certain parts of the page fast and your students are still going to be waiting for the video stream to load and, you know, to connect, quizzes. Jesse Friedman: We brought this up in a previous podcast episode when we were talking about concurrent users, concurrent logged-in users with, within higher education especially around multisite as well. That can create a large burden on the site. Christos Paloukas: Yeah, that's when the actual systems play a role. So what happens there is, we do the load testing. We come up with the results. We definitely need expectations first from the customer, like. Okay. How many concurrent users do you expect? How many people do you expect logging into your website and what do they do? They click on a certain course, follow that course through, maybe there's 10 different pages on that course. Maybe there's a live stream, maybe there's just videos, and then there's usually a quiz like either intermittent quizzes or a quiz at the end. Quizzes are heavy, you know, they've got that POST request. And if you've got a lot of students submitting a quiz, you might end up with, you might even end up with timeouts. Like at Pressidium, our timeout is, our default timeout is at two minutes, you might have students pushing the button then waiting, waiting, waiting, waiting, waiting. And the request is never, it's lost, it never makes it to the backend. That's a bad situation. Especially for the students on the other side of things, if you see it practically, if it's like some sort of university that I don't know you've paid for and you're like locked in, you're probably not gonna lose that student slash customer, I mean, to be frank, but they're not gonna have a good time. Jesse Friedman: You're gonna get, you're gonna be pissing people off, Christos Paloukas: Yeah. Jesse Friedman: Frustrating them, and they're gonna be annoyed. They're trying to get something done or get a project turned in on time and they're against the clock. And Christos Paloukas: Yeah, Jesse Friedman: It creates that stress for the end customer. We started talking about segmenting customers. You're talking about how you have a number of enterprise customers, but you also have a larger number of smaller websites or people who are doing this on their own. Conclusion Jesse Friedman: So what I wanna get into is talking about how do you lay a foundation that works for everybody, and all that. Because I think that's super important as we think about the way in which managed hosting applies performance layers to a website and optimizing a website. But we are at time for this episode, so I do wanna just take a quick break. Thank you so much for being on, and we'll be back in the next episode. Christos Paloukas: Yeah, it was a pleasure.

Open the transcript file

Hosts and Guests

Hosts

  • Jesse Friedman Host Feed-stated

Guests

  • Christos Paloukas Guest Feed-stated

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Revisited: What's the Future of the Open Web with AI-Powered Content?

    Nov 10, 202636 minAudio

    Are we all just orchestrators now? Jesse Friedman and Chris Coyier discuss the shift from writing code to conducting AI agents, what that means for the future of websites, and why the web might end up bigger and...

    Credit the people on this episode

  2. The Creator Economy Is Built on Rented Land. Hosts and Agencies, Go Knock on the Door. (Video Essay by Jesse Friedman)

    Oct 27, 20269 minAudio

    Hosting companies have an opportunity to serve creators who depend on platforms they cannot control. Jesse Friedman of WP Cloud recommends selling audience connections rather than infrastructure specs. He shares...

    Credit the people on this episode

  3. What Role Should Hosts Play in Simplifying the WordPress Experience? (with Andrew Palmer)

    Oct 6, 202636 minAudio

    Low hosting prices can come at the cost of a better customer experience. Jesse Friedman and WordPress advocate Andrew Palmer discuss why higher prices need to deliver simpler setup and meaningful services. They explore...

    Credit the people on this episode

  4. What Should Hosting Companies Do to Reduce Customer Churn? (with Andrew Palmer)

    Sep 22, 202631 minAudio

    WordPress onboarding is still too hard, and Andrew Palmer thinks hosts can fix it. He and Jesse explore why hosting companies should compete on the WordPress experience instead of increasing stats like PHP workers and...

    Credit the people on this episode

  5. How Does Accessibility Improve Conversion Rates for WordPress Sites? (with Anne-Mieke Bovelett)

    Sep 15, 202635 minAudio

    Fear-based messaging does not work for accessibility any more than it does for security. Anne-Mieke Bovelett and Jesse Friedman explain why a caring, positive approach gets better results when helping people build...

    Credit the people on this episode