Eric Karkovack (00:00) Hi everyone and welcome to the WP Minute. I'm Eric Karkovack. So today's topic is timely. The holiday shopping season is upon us and that means preparing our WooCommerce shops for a surge of traffic and orders. A high performing website is crucial to making sales and satisfying customers. But what does WooCommerce performance look like? What server configurations do it best? Today's guest will shed some light on the subject. Matt Schwartz is founder and CEO of Inspry, and he joins us to discuss his company's recently updated WooCommerce hosting benchmark results. Matt, welcome back to the WP Minute. Matt Schwartz (00:37) Thank you, Eric. I'm super excited to be back. Eric Karkovack (00:39) Yeah, we we talked a little bit earlier this year about about a SaaS that you run. and it's good to have you back. And so tell us a little bit about the background of the WooCommerce hosting benchmarks. Like why did you do this? Why did you as an agency owner, like what prompted you to do this? Matt Schwartz (00:57) Definitely, yeah, that's great question, Eric. to kick things off. so it you know, i short, I probably went into a little bit of a psychosis to to get to figure out the you know performance for this particular client. but the the longer story is, yeah, I was working with a client. they have a WooCommerce store. this store essentially they sell electronics that they manufacture, both consumers and businesses. In terms of traffic, I would say like they're not a huge high load. they do maybe between thirty and fifty thousand, you know visitors a month. which is in a small store, but it's not like massive, right? But they they do have a lot of active subscribers because as part of their product, you can subscribe to it has like a subscription portion to it. So, you know, they have around at the time fifteen thousand Woo subscriptions and they had like Automate Woo running. they had Elementor, all these different plugins, and they already came to us with the site pre-built. We did an audit and we inherited the site essentially to help build new functionality into it. And so with that, one issue that kept coming up was, you know, the owner was like, hey, we do a ton of phone orders. We, you know, are constantly in the back end of our site. And you know, our team, every second that we're waiting for, you know, orders to load and to search for orders and look for products on the back end that really is affecting our sales and just how many people I can have within the office. So I think this isn't an uncommon situation with with a lot of clients. but so yeah, that's kind of where it started was he said, Hey, figure this out. I'll give you free reign. he said, you know, I'll even help pay for, you know, some of this. So we we first did some software optimization obviously within Woo. we did some different things there, HPAS, you know, PHP upgrades, some indexes to the database, some query changes at his current host. but unfortunately it wasn't moving the needle as much as he really wanted it. He was like, I really want this thing under a second when you search. That's what I want. which is pretty fast. yeah, right. Like a lot of a lot of woo sites don't need that, I will say. but I think it was a really neat example of how you kind of hone in on your clients' needs. as an agency and that's where you kind of base the process of finding the correct solution, whether it's in software or on the host or or you know, a hybrid of both. Eric Karkovack (03:32) Yeah, that's it it's one of those things where the client has a a specific need and then you have to go and s search and search around and and find something that will do that. Now so you went around to these different hosts, right, a and wanted to to create this benchmark. How did you choose hosts for this and were they aware of what you were doing, or did you just go and buy hosting accounts and say, let's let's see what this does? Matt Schwartz (03:58) Yeah, so originally this was not going to be a you know a public benchmark by any means. and we have worked with a lot of different hosts over the years. You know, INSPRI, our agency, is partners with some of these hosts, just an FYI. but you know, we're not really in the marketing online space. It's more just when we work with a client, we may refer them to a host, right? And that's how we're partner. Eric Karkovack (04:23) Yeah. Matt Schwartz (04:24) but essentially even their the The hosts they were with, you know, we were good partners with, but we just couldn't find a good solution. And the client was like, okay, we need to, you know, look at other hosting options. And we agreed. We said, yeah, let's do some tests. And so we essentially obviously looked at some hosts that we'd used in the past. Actually, the previous year, our agency, we did not release these public results, but if anyone is ever curious, we did some load testing. or Woo, essentially where you're getting huge spikes of traffic. you know, you might have five thousand c concurrent users or ten thousand concurrent users. checking out which on the technical side typically is not easy to cache right so that uses all the resources of the server. So anyways, we we had already done some benchmarking in that capacity. So we knew some hosts we wanted to start with. but we were also interested in some of the different options out there that we hadn't explored especially like you know there are panels now like Xcloud, right? That you can attach onto more of either a bare bones or a managed raw server and we knew a lot of agencies doing that to save costs and get more raw performance. So you'll see a lot of these benchmarks we used used XCloud just and there's a few other ones out there. Not they're not the only one, but and then we combine them with you know different servers like liquid web, et cetera, to see how though how those would go. So essentially our goal was we want to host ideally that is fully managed because that's where they were coming from. We don't want to be doing OS updates. So even the XCloud examples we gave all were hosts where they were managing at least the OS, right? At a minimum, because we did not want to do that. Of course you pay more when you add that, but you can add that to different providers. Or they might be more of a, you know, a VPS or cloud managed WordPress host, which is still our highest preference, at least at our agency. That may not be the same for every agency, but we consider ourselves fairly technical, but ultimately, you know, we're not we don't need root access typically to the server. We're running, Woo and WordPress. And you know, we can get what we need with like SSH. And that's, you know, plenty. So that's really how we landed on those hosts. It was looking at, you know, hosts we either knew or just researching online other providers that people specifically called out for WooCommerce hosting with things like backend page speed and just overall page speed with Woo. So we really were just looking through Facebook agency groups. Reddit, you know, all the normal places that you could to put that together. Eric Karkovack (07:05) So Let's talk a little bit about the metrics that you you chose for this. Now I know you we always talk about the front end because everybody wants the front end to load quickly when a customer comes to the site, but you are also measuring back end stuff, which as you mentioned is important when you've got you know a clu a customer on the line and you've got to look up an order or place an order through that that back end. what did you measure here and you know was there anything that you know Maybe you weren't expecting to measure that you came to, at some point in the process and said, Yeah, we should probably measure this too. Matt Schwartz (07:41) Yeah, yeah. So like you said, I think everyone focuses heavily on front end speeds. when you go online, of course you can see all these benchmarks and they're just demoing this with like demo, you know, page builder sites or whatever it is, woo. And they're usually very Clean copies, where we wanted to focus on a real client site. So all the messiness that, you know, we we tried to clean up as much as we can, but you know, the client also wanted some plugins that are heavy by design, not that they're bad plugins, but certain page builders, certain security plugins, etc., that you know, process a lot. So that was kind of the first part. and then outside of, you know, just using a I would say a a real outside of using like a r a real site, we also wanted to look at the back end, as you mentioned. And search specifically, because that was super important for this client. And that's kind of was saying you want to focus on what I think your client ultimately needs from the business sense. And that was his number one thing was like search back end. And there were some other tools we could have used. For example, we we tried some plugins, we tried some queries, et cetera, to clean it up, but ultimately really the only other option we had found outside of just using. better hardware, better hosting, was you know, potentially using like a second database server, which is more complicated. Or, you know, there are some services out there now where you can offboard old orders actually to like a SAS in Woo, which is kind of cool. so then when you're searching in Woo, it's actually searching like a optimized non-woo database somewhere else, which I think is a brilliant idea. it's kind of like what I think what ShortCart might do. something similar. But essentially you know, we had those kind of you know worked out in our mind. Later on we added value as the fourth one because cost, right? What is your dollar average for performance essentially with all these metrics? So that actually went into all the metrics as well in terms of how much would you be paying a month for this? and so those four we basically split equally ultimately with front end, back end search and value. I will say like front end was not the original goal. But it just made sense at that point because people care about that. And we still needed to make sure when our client moved that they were still happy with their front-end speed, right? even though they weren't doing like massive amounts of front end traffic, still. So we split it into those four, and and then that basically set up kind of the benchmark that we worked with to you know start the process. And it was somewhat based a little bit on those previous, like I said, low testing benchmarks. We'd done some of the before but this was like you know the first time a client really had specified exactly what they wanted and they wanted to see all the results and all the metrics. So it was pretty neat. Eric Karkovack (10:33) So what what did you take from the results? Like 'cause we can look at and and we see the hosts. Some of them did better in other areas. like was there anything that surprised you when you got these results? Matt Schwartz (10:47) Yeah, so I think the front end results honestly were not as not very surprising. just because I think so much time and effort is spent on the front end that a lot of hosts, you know, they are using CDNs and caching and they pre-bake a lot of this in for you, right? but I think that definitely back end speeds were not something I think hosts either had as many optimizations built in for or had the amount of support you might expect when let's say the the back end slow because when we even did this, we had a pre-written script that we give all the hosts. So we did tell the host to answer your earlier question, we did tell the host we are benchmarking and testing you against other hosts for a client, which at the time was true. Again, it wasn't public. We did sign up ourselves. We paid for all of these ourselves except for Kinsta, which ServeBolt offered this year to actually help essentially pay for for that test, but all the rest our agency paid for, just out of pocket because we wanted to really know the results, right? So ultimately though, the back end was super surprising. The search was super, super surprising, even with things like Redis on, because pretty much all the host, actually every host that we tested had Redis or Memcache in place. but definitely the amount of I would say service and support you got varied and things like PHP workers didn't really make a difference, which makes sense because PHP workers just allow more executions really. but if the bottleneck is The database, which is often the case with Woo's backend, with Woo queries, even with HPOS, it doesn't really help there. So that was kind of an interesting thing that we learned. And, you know, we had hosts even tell us that. Like you can upgrade to the next plan up with us, especially the cloud-based ones, but it's not gonna help you essentially, because adding more PHP workers is not actually going to help with one execution that's taking longer. and you might say, okay. Well, it sounds like you had essentially like a a query or bottleneck issue. And some of that probably is true, right? But I do think some of it was because of the plugins, et cetera, that the client wanted to use. I think it's also because Woo is continuing to improve, you know, their queries. even in, you know, later versions of Woo, I know they did some work around that. But essentially, even natively, there are some limitations around that. But it even with Woo though, you can still improve it both in software and through your host, your hardware clearly, as our benchmarks show. Eric Karkovack (13:23) Yeah, we focus so much on things like memory, storage, CPU cycles, but the database is such an important part. I mean, when you're talking about a site that has, you know, a hundred thousand products or whatever it is The database is gonna make a big difference, right? So did you find okay, so you've had Redis or Memcache on there, did you find like a specific host that blew you away in in database performance compared to the others? Matt Schwartz (13:52) Yeah, so I will say that, on the back end and on the database, Bigscoots did really well specifically. and I think it was a mix of they have possibly and again, I'm not gonna pretend like I'm in a hosting expert here. You know, I'm an agency owner. But that was kind of the point of this whole thing is, you know, I think I have more knowledge than some agencies and definitely more than a lot of site owners. And our goal here was like let's pret let's do this based on if I was really the end customer and I want this set up how the host says. So when we went to Bigscoots they actually were very kind of directly, you know, hands on. They did additional indexing that other hosts didn't do, even when I went back to the other hosts and say, hey, do you have any, you know, indexing want after, you know, Bigscoots offered to do it. I was like, well I need to be consistent, so I'll go back to the other host and ask. But they did that. and from my understanding, just how they they work with their setup, they do have a little bit more control over, you know, how they allocate the resources, etc. It also was a More expensive plan than practically every plan that was in this benchmarking. So, you know, you have to keep all that in mind, but they, you know, they out of the box got it under a second with searching and everything the first try, which was pretty neat. and then I would say second, which, you know, again is in the benchmarking is Servebolt they Without any real hands-on need you know, work, they were able to essentially achieve a fairly similar result, almost the same, on the back end, which was super surprising because again, I didn't know them very well. I hadn't really, besides just, you know, offering to to do some basic optimizations which they did. they weren't super deep on the site, but out of the box it it actually did really well with Servebolt as well. So it's pretty neat. Eric Karkovack (15:43) Yeah, so I do want to mention too that both Bigscoots and Servebolt are Foundation Plus sponsors of The WP Minute So we thank them for that. but I always want to make sure we disclose those relationships. And and you know we have other hosts that were on your benchmark, like I know Pressable was one of them. They're they're on that list. so I you mentioned cost too, which I I think's interesting because you had a pretty good range of of costs. You had from you know, like the thirty dollar of month plan to the $300 and some dollar a month plan. But cost doesn't necessarily correlate to performance, does it? I mean in some cases maybe it does, but for for certain things. But what what did you find there? Matt Schwartz (16:24) Yeah, so the we had a a good range, partly because we went to the the companies and said, Hey, like, help us figure out which which plan's the right fit for this site. And we gave all the stats on the site, right? We gave the analytics, we gave them these are the metrics we're gonna be testing you on. So we gave them all that up front. but it was interesting because yes, like there was some I will say some on the value side, like rocket.net, they did very, very well for what you were paying, $30 essentially a month. Eric Karkovack (16:58) Yeah. Matt Schwartz (16:59) so you know, from that perspective, it can go either way. You know, some of the other hosts were more expensive, but they were more like a middle of the pack, you could say. Eric Karkovack (17:09) Yeah. Matt Schwartz (17:10) So it's not always you get what you pay for. I think obviously there's probably some truth to that, especially in like the shared hosting. You know, you're talking like the five, ten, twenty dollar plans. that that is somewhat true, I think, at least in my experience. But definitely once you start to go past that, I think it comes down to less about the dollars, but more about again, what is most important to you on this website. Because with all of these hosts, all of them. there were caveats to every single one. even the best performing ones, there are things that are not covered in this benchmark that are not just performance, right? It could be their dashboard. It could be what you can do, you know, with their system. You know, each of the hosts definitely has benefits. but there are trade-offs. And so I think that's why it's all so important, just like with any agency project, you start with what is the business problem and you work backwards into the technical solution. And I think with the cost side, yes, you have to consider cost, but I think it's more important to consider what are you technically trying to fix or resolve. and then you'll see, at least in our example, that you know, the cost doesn't always correlate with that particular item based on, you know, for example, again, Servebolt will tell you this themselves that they overall did the best in all of the in all of the hosting examples in terms of they came second in every place, which technically made them first, right? Because no one else essentially scored that high across all of them. but you know with that at especially at the time, I wasn't familiar as much with Servebolt Now I'm definitely more familiar. Our client wasn't familiar with them. and we weren't ready to take the plunge at that point with this particular client. But they they still did really well. And I think that that's just another example. You have to think outside of all the variables that are important to you. It's not all benchmarks essentially. It's not all cost, right? you've got to look at every part of that. It's the support The dashboard, Eric Karkovack (19:08) Yeah, that's a big one, right? Matt Schwartz (19:10) all of those. Yeah. And I will say, with most of this we tested, the support was pretty good in general. I was pretty impressed with, you know, of course, I have some favorites around support, but I do think pretty much everyone did pretty well on the support side in terms of, you know, trying to help us set this up, at least with it, you know, the knowledge and the infrastructure they had. Eric Karkovack (19:30) Well that's good to hear because the support is like one of those things that you're gonna need. you're hoping not to need it, but you're gonna need it at some point. You're gonna have questions, you might run into issues, you need to know certain things. So it's good to hear that everybody's responsive. because even if you are only paying 30 bucks a month, you still want to get answers, right? And so that's a big deal. so what would you say like if you're looking at this from the fifty thousand foot view, what is this what are these benchmark results? Matt Schwartz (19:58) I think it shows that there are very good hosts for Woo out there. I think there was this kind of myth for a long time that you can't have good Woo hosting essentially, and that ho Woo's gonna be slow and all of this, right? and I think this shows to me that there's multiple hosts that are serving Woo very, very well. and again, there's obviously benefits to each one, but there are out of the box answers. that, you know, if if You're struggling with either scalability or even load testing. there are options now that are Woo optimized. And I think that's fantastic. I think it it really I'm hoping that will show, you know, some trust in Woo. And obviously Woo has been again optimizing their own infrastructure, just on the software side too, within WordPress. But yeah, I think that that's probably one of my biggest takeaways. And then I would say Another takeaway, which you mentioned earlier, like with support, is that you know we still host a ton of sites with Pressable right? And I've been very happy with Pressable for a lot of our websites. Their support's been great. But you don't always have to use the same host for every project, right? And I don't think there's anything wrong with that. There's obviously some benefits to doing that in terms of the systems better, but it's a lot Eric Karkovack (21:18) Yeah. Matt Schwartz (21:18) like software in general. You don't have to use WordPress for everything, right? I don't need to build my project management software in WordPress. You can, but you don't have to, right? There are there's benefits to not doing that sometimes. So I think it's similar with hosts. you don't always have to use the same host. there are is definitely hosts that it excel at certain projects better than others. And I think that's the right approach. And that's really what I got out of it. And it yeah, it just kinda has strengthened how when our agency looks at hosting our client sites now, that's really the approach we always take. Eric Karkovack (21:52) Well that's great. it it's good to know too that's out of out of these options that you you picked, and I should mention too that Kinstas Kinsta had their test done like a I what about a year later than everybody else because of the the Servebolt request and Matt Schwartz (22:08) Correct, yeah. So so and just to give some transparency, because I want you know, users to know where we sit with this is we did all this benchmarking internally at our agency. Essentially I ran into ServeBolt at PressConf, which y'all might be familiar with, and told them, hey, yeah, I benchmarked you guys. I knew nothing about y'all, but y'all did great. And they said, well, that's fantastic. You know, I I I I would love to see those benchmarks. and so, you know, we we were able to with their sponsorship test Kinsta as well because they originally just Terms of budget and time last year we weren't able to test them. So we added Kinsta as well into the mix with their sponsorship. But that ultimately I think really showed character in Servebolt to me because they didn't try to influence anything we did here. this was all done, you know, in by us, and that's exactly what I kind of just said. Is with this particular client, they didn't even pick Servebolt but at the same time, I think that by including Kinsta and you know, Servebolt and and myself, we were able to show another example out there and Sir and Kinsta did very well on the front end, but to your point, it was done a year later later, so I do want to be transparent with the audience that it was on newer Kinsta infrastructure that when they moved to a new Oracle cloud infrastructure. And it was on Woo 10, not 9.9, which is actually what these were all hosted on. So there were some differences there. So do keep all of that in mind. But that's essentially, you know, I still think there's some value there in in seeing how that works, especially because Kinsta, if Kinsta wasn't included, you know, one of the other hosts actually would have been Faster, right? So it showed just an example of how yes, depending on what you're doing, a certain host might be a better fit. And even over time, you may want to consider where you host, right? You may want to s look at other hosts and do testing later on because as hosts improve their infrastructure, they're of course gonna throw a bunch of marketing at you. But if you go do your own tests with like we did with Query Monitor and GT Metrics, you can really do, I think, the best service to your your clients or your Users on your website by doing a check-in, you know, every couple years, especially if it's a big site, and seeing what other options are out there. Eric Karkovack (24:26) That's a good point because yeah, th these these hosting companies obviously do not stay stagnant for very long. I mean that there's always new hardware and new software solutions coming out. So if you're a freelancer, if you're an agency, you've got clients using WooCommerce, we're gonna put a link to this in the show notes. You can check out the benchmarks. And Matt, is there anything else you wanna let people know about about this these tests? I mean any future plans to update them? Matt Schwartz (24:52) So yeah, I mean right now, probably not only because this is a as a, you know, agency we don't have necessarily the time to, but I would love to see more of these sort of benchmarks made by agencies, you know, so I encourage other agencies to, you know, if you can, you don't have to do like twelve we we did, right? But I do think seeing more real world sites like this is gonna be very helpful, I think, for other agencies. and you know, I would love as an agency to see more examples of that. Like if those had been there, obviously that's something that I would have considered as I went through this. And that's why I also tried to kind of cover everything that we did specifically. So if another agency wants to kind of, you know, use our process as an example, whether you're using it just for one client or you want to publish it online you'll be able to really, you know, do that I think on your own. And and then one last thing I do want to mention before I forget is When it comes to the CDNs, you know, on the front end, obviously CDNs made quite a big difference. But as you would expect, on the back end, not so much. because it's not as cache So like ServerBolt didn't even have a CDN when I tested it, because that plan that I was on, which was a a free plan. I basically had messaged them and say, Hey, I'm testing your system. And again, it wasn't for any public benchmarking. They're like, here's a free plan. And it didn't actually have the CDN enabled at all. but I think that that's what it showed. Servebolt still did really well without the C D N on the back end and still quite well on the front end. But I I think that is something to keep in mind. A CDN is not gonna fix all your, your problems on the back end essentially. You've gotta you gotta dig in a little deeper. Eric Karkovack (26:37) Yeah, it's definitely more complicated than just throwing Cloudflare on on the site. you've got to look at at every angle. Matt, I appreciate you being on today. Where can people connect with you? Matt Schwartz (26:48) Yeah, so I'm definitely on LinkedIn, check me out. Our agency in Inspry check out checkview.io if you have Woo or WordPress sites that you know you're worried about your forums or checkout breaking and not being aware. but definitely mostly LinkedIn. I'm also in in the table or in the admin bar. so definitely if you're not in that agency group, it's a fantastic place of knowledge as well. Eric Karkovack (27:13) Awesome. Well, thank you for being with us today, and thanks to everyone for watching and listening. Visit us over at the WPMinute.com slash subscribe. Get our newsletter, support the work that we do, and we will also put a a link in the show notes to Matt's previous episode where we talk about CheckView and what that does, because it's an awesome tool. thanks everybody and we will see you again next time.