Episode Transcript
[00:00:11] Speaker A: Welcome to Arbery Digital experiences, episode seven.
And today's subject and topic is going to be delivery to folks inside the borders of mainland China. And so we're joined here today. We've got myself, Ted Reeves from Arbor Digital, as well as Dwayne Hild, who's an AM engineer from Arbor Digital. And we're joined by Jody Chan from Chinafi. And Jody, would you like to introduce yourself?
[00:00:38] Speaker B: Absolutely. My name is Jody Chan, and I am the senior vp of product and strategic partnerships at Chinafi. It's a very long winded way of saying I get to work very closely with our partnerships teams, including you guys, hopefully in the future, as well as our product team in terms of the development of the product.
[00:00:58] Speaker A: Going to, we're going to dive into what your product is and does and so forth. But what I wanted to do first, before we even get into that, because there's a lot of folks who may not know even the framing for the problem of why does China have to be, or why does any specific geography of China have to be treated separately from the rest of the Internet? Because you don't say, I need to make my site optimized for Spain from a network perspective, or I need to make for whatever, Switzerland or something like that. So what makes China different? And what are the content limitations? What are the network limitations? What do people have to worry about and be concerned about when it comes to just, they've got a translated language site hosted on whatever platform they have? What do they have to be aware of? Do you want to give us a little bit of an overview just from what you've seen, Jody?
[00:01:49] Speaker B: For sure. Well, I personally find this topic endlessly fascinating because I get the pleasure of talking with folks from all over the world and from, I guess, various decision making capabilities at different companies and even those that have been in the industry for most of their careers, actually, oftentimes find that China in terms of web performance is a big black hole, and it's something that's very nuanced and very difficult to comprehend. What most people will have heard of, and I'm guessing some of the listeners on this podcast will have heard of, is the very broad term of the great firewall of China. Right now, that broad term is great for conversation because you say, why doesn't it work? Well, it's the great firewall of China. But what does that actually mean? And when we break that down, actually, one of the key things to highlight is how we define performance. Right? So the way that we define performance when a website works or doesn't work, can largely be broken down into two components. The first component is obviously speed. The website has to load within a reasonable amount of time. This isn't even going into specific metrics like time to first byte. The new metric of Inp, things like that, just has to load within a reasonable time. The second component, in terms of determining when a website works or not, is functionality. And so what a lot of people don't realize is when a website, no matter how well designed or however it's built, is loaded from China, unless it's been optimized or tweaked in some way, it's more often than not loading incredibly slowly, if not timing out. And to my second point, it does not function. So you'll be missing videos all of a sudden, fonts won't be working. Maybe you won't get any analytics firing at all. Maybe from a user perspective, you're trying to click a button, but nothing ever happens. So that sort of thing can happen with the best performing websites out there. And so to highlight and just give the background to this conversation is that no matter how you build your website, no matter how well designed it is, no matter whether you're using the best practices out there, it's more likely than not going to encounter some type of problem for visitors coming from China. And the great firewall of it's, I would say it's both a political and a technical rooted problem. Right. You can't deny that there is a degree of regulation and censorship involved. Now, we're not here to kind of give opinions on that. It's kind of like when you enter someone's house, you kind of play by their rules. And this is sort of the analogy that I would put as it relates to the web in China. And so now, as a web performance person or a product person, the role is just to be aware that there are visitors in China trying to access your websites. And even if the content itself is not incendiary or problematic, it's still going to run into some sort of technical issues. And that's what we're here to talk.
[00:04:43] Speaker A: As. And you and I were talking about this a bit beforehand, and I think that one mistake that you could even get into, you could probably go all day talking about what mistakes people make when trying to design a for China audience website. But it isn't just if you even have, like, a lot of websites aren't translated into either simplified or traditional. You've got, even if you've got an english language site, if you would like to have it viewed by people inside China, even if it's just all English, then this is still an issue.
[00:05:23] Speaker B: Absolutely. In fact, actually, I work with so many companies, and their purpose of making their website work for visitors in China really is for credibility purposes. The content itself might still be in English, but you have partners, you have potential investors. A lot of the times it's investor portals, actually, and they're still primarily in English. But it doesn't mean that your visitors in China have access to that information or should be limited, I guess, on the converse, to access to that information.
[00:05:54] Speaker A: Yeah, that makes sense.
Also, just so folks understand this, too, specifically, when you're talking about what, geographies have a limitation on this. We're talking about specifically inside China itself and excluding what?
[00:06:12] Speaker B: Excluding what? Do you mind?
[00:06:16] Speaker A: So this is China itself, not Macau, right. Not Hong Kong. And in terms of just chinese speaking regions, we're also talking not Taiwan. Obviously, it's a different political status.
[00:06:32] Speaker B: Mainland China, generally, calling it mainland China is the easiest way, not any special administrative regions or other geographies, but specifically from mainland China.
[00:06:42] Speaker A: Good.
Cool. All right, good.
Because you talked to a lot of companies, and you've also dealt with a bunch of ones that use am as well, which is a lot of the reason why we're talking, because am gets super complex, and it just gets more complex when you're trying to host in China. But what do you see? If you're going to summarize a couple of things that you see in terms of approaches that you wish that people knew about or wish that directions that you wish companies would be taking when approaching a chinese audience, what would you say to something like that?
[00:07:23] Speaker B: Well, the first thing I would say is that it's relevant to almost everyone. I think one of the first things that I hear when someone has not realized that there is a problem with their website in China is that that's not relevant to me. And when we're talking about companies, if you are talking to the highest level stakeholders, they will say, no, that's incorrect. China is absolutely important to us to some degree, no matter what kind of company or industry, because it's the second largest market in the world. And as the second largest market in the world, you're not just talking about consumers there. You're talking about people who travel there. You're talking about anyone who's within those borders will encounter some sort of problem accessing information from what you're trying to share with the world. And so I think the first thing I would share is it's relevant even if you think it's not in the beginning. And the second thing that I'd probably share is that there are a lot more barriers than a very experienced developer might realize when it comes to China. So when you kind of solve a web performance problem, most of the time you kind of pull from an arsenal of tools that you've kind of used on every single website. You might move the hosting server closer to the end user, reduce latency that way. You might bolt on a content delivery network, you might minify js, you might try to optimize your images, things like that. There's an arsenal of tools that you can use in most cases outside of China, that actually help you achieve your goals. And when you achieve those goals, most of the time you're talking about taking a website, let's just say in its worst case scenario, outside of China, let's just say 10 seconds, ten to 15 seconds. And you want to get it down to five, three, two. When we're talking about China here, we're talking about taking a website that times out, or at best is 15 to 30 seconds loading and incompletely loading, might I say, to something that loads within a reasonable amount of time, maybe five to 8 seconds. So the gaps that we're talking about here are extremely, extremely large. And those arsenal of tools, while helpful in optimizing the original website and obviously reducing latency to some degree, we're not addressing the core issue of the problem here that arsenal applies but does not solve. And so when it comes to developers and just purely awareness, that's the second thing I would highlight is your arsenal. Tools are great, keep working with those. But be aware that that's not the complete set of tools that you need when tackling China. And so, yeah, first thing, relevance. Second thing is, you know what you don't know. I guess that's one way of putting it.
[00:09:59] Speaker A: Yeah.
[00:10:00] Speaker B: And the third way is really just to understand that there are solutions to this. Right. A lot of the times what ends up happening is that upon realizing that the tools on hand are not able to solve a problem, you might go to Google and start searching for ways to address this issue in China. And you'll probably come to the conclusion that there's essentially only two routes to optimizing and making a website work in China. The first route is essentially to build a more or less simplified website. And it's not to take it in an offensive way. Maybe it looks like something from the 90s because it's not offensive at all, because it works. Right.
[00:10:37] Speaker A: We're talking the workforce in the talking.
[00:10:40] Speaker B: About, let's not age all of ourselves here.
I'm referencing something that maybe some folks listening will not be able to reference for the most part. But let's just say it's a very simplistic rebuilt custom website that you then host in China. And to host in China, you have to have a China business entity. You would need something called an ICP filing or an ICP license, depending on the type of website, PSB filing. More often than not, you need an onshore developer to manage the. So, you know, these things can take anywhere from two weeks to two years and headcount maybe, I don't know. I would say above one in general, irrespective of what you're investing here. And then more often than not, you also end up encountering issues of language barriers because some of the tools that you would be required to use are also purely available in simplified Chinese or the support is limited in English. We've encountered that as well, where we've had to act as a middleman between different vendors and things like that. And so I would say the third way is there are solutions to address this problem, but they're not always straightforward. So that's the first route, right. That's the route that might make sense for companies that are, by regulation, required to host their information in China. And by all means, they absolutely should. Right. I mean, follow the law is kind of our stance across the, you know, that's one way, but the other way really is. And that's kind of where we'll transition to Chinafi. Because what we're trying to address here is Chinafi works with websites that are both onshore and offshore. And what we're essentially doing is identifying components of the page that might be causing you those performance issues, either because they're blocked and need a replacement or they're incredibly slow because let's just say it's hosted on a third party domain and that domain is placed very far away from China. But the way that they're doing those calls, scripts, the ways they're written, et cetera, just amplify the problem. And so to my earlier points about the three things hope, you know, someone takes away is that this matters to them. Two is that you need a specific set of tools when it comes to China, and it's nothing to do with how experienced you are as an individual. And the third point is there are solutions to this, but the best way to do it is to partner with someone who's actually done it. And to be able to go in with the objective that you want to maintain and have that level of creative freedom and technical freedom that you do right now, for example, with you guys building a beautiful, dynamic website, you want to be able to do that and deliver that to your visitors in China. That is the objective. And how do we do that in a technical feasible.
[00:13:28] Speaker A: Absolutely. And so I wanted to kind of talk about that last point because I think that last point of, just so you know, something can actually be done about this.
That point I think stops is something that a lot of companies don't realize because they think that, because they're already at their limit when it comes to hosting the rest of everything else. And when you start talking about solutions, potential large scale solutions around really localizing support in China, then you see their eyes start to roll back in their head and they're going to just pass out because it's a lot, and especially when you talk about AM and I wanted to actually just talk a little bit about kind of framing the problem from an infrastructure perspective because both Duane and I have hosted things that are running near China, running around China, or have specifically avoided China because of cost constraints or technical constraints. So for one, so if you're running Adobe experience manager, a lot of people run it on premise. If you're running it on premise and you do at least have the flexibility of sticking that gear where you want or where you can afford to, a lot of times that might have been at an on premise data center, at a third party, at a co located place like Equinix or rackspace or something like that, or they're in AWS or Google or something like that. And then when you say, okay, well, we want also to serve China, then you, we, so you're going to have to provision something out there. So you need to have a data center provider that is in.
So sometimes that comes with learning a brand new cloud provider like, oh, I'm certified on AWS, but what about Ollie cloud? I don't know anything about Ollie Cloud. Should I trust like all that sort of stuff, you just kind of go, it's a completely new world.
But then if you go and you're already in a hosted platform like Adobe AM managed services, which does not actually technically support China, the umbrella marketing term supports China, but the in China solution is an entirely separate solution that has nothing to do really with the rest of the world because it's not hosted on the same platform, it doesn't use cloud manager, it doesn't use the same CI CD framework it doesn't use all the different frameworks are totally different. So running that deployment, you basically need two completely different deployments already.
But let's say you moved on from all that. You're like, okay, I'm on the new step. I'm on aim as a cloud service. Aim as a cloud service specifically only runs close to China. The closest you can get is Singapore and Tokyo. But you can't get into China.
That's as close as you get. So you don't actually get on continent.
And there's no plans to put it inside the borders either because it would just be technically unfeasible with the tooling that's available. The new AM edge delivery services also is the same way. Both of those use fastly as their edge side CDN. At CDN fastly doesn't have any points of presence inside.
Yeah, and it gets even more interesting in that, well, let's say I want to just go and put something there, but I'm going to use all the same monitoring tooling and so forth that I use for everything else.
A lot of the big monitoring providers also don't work inside mainland China. For example, the one that Adobe uses for all of its hosted services, in which a lot of us know and love, which is new relic. New relic doesn't operate in China whatsoever. You have zero operations in. So, and that makes it really tough to do things like measure performance, which comes back to your point of like, you kind of need a team and you need people who are used to monitoring this kind of stuff because a lot of times you go, how bad is it?
Okay, good. So you're saying my site, okay, I got a complaint. Somebody said they waited 30 seconds for my site to just not load.
Is it that bad for everybody? But you don't know because all your tooling that actually says, and maybe you run a web page test and the web page test is not that bad. And you're like, so it says it's not that bad, but I think it's that bad. Then it becomes really troubling to be able to do this. But then back to this earlier thing of like, okay, good, maybe we should put stuff in China.
So individual AM licenses are very expensive. Let's say you're already running whatever. You maxed out your budget with three publishers or something like that to say, okay, good, I'm going to buy another one for China. Getting that through finance and justifying that, you have to justify the cost of another, let's just say between the license and the hosting and all that kind of stuff. Just say, just for licenses and gear, you're going to spend $150,000 or something like that, and then you get to spend the money on a guide to put it in there and test it. And it's going to be multiple people, just like you said. And then it's like all this expense, and most people go back to just passing out because that's easier because you just go, good. Well, I can't do anything about this. I can't reach people in China.
I'm so sorry. I really wish I could, but I just can't.
That's true.
[00:18:49] Speaker B: That's the most heartbreaking, you actually trying and going through all the motions. And you'd be surprised how many people we've actually encountered who've done that, who really run it, maybe to the 90% of this path that they've been told that they should be on, and to come back and say it's still not working.
[00:19:07] Speaker A: That's right. And even things that are close sometimes, sometimes when you're running the infrastructure, like Dwayne, you had things that are running in Hong Kong that sometimes, just for absolutely no reason whatsoever, you'd have weird replication lags that just, things stop and they shouldn't. You know what? Right. Like, you're having all kinds of.
[00:19:25] Speaker C: It was not only replication lags, it was also issues with, the way I would describe it is the edge. Users on the edge of the continent were getting pretty good response times. They were able to download pdfs pretty regularly like most end users would. But it seemed the further inland you went, kind of to the point earlier, there wasn't a whole lot of pops further inland. And it just seemed that one day somebody in x area would have an issue, and then the next week they would report that it's working just fine for them again. So it was very inconsistent in behavior just out the gate.
[00:20:04] Speaker A: Yeah, it's infuriating to deal with sometimes.
I guess that may be an adequate segue to just. Okay, so your solution, Jody, can you just talk me through a little bit? Like how just basic nuts and bolts, visualize me some, like a block diagram in my head of how this would bolt?
Actually, let me give you a scenario. Okay. So I've got an AM cloud service site.
So with the Am cloud service, you get to pick a place where it's hosted. So let's just say I picked Ohio. So it's North America, right? So I thought that was going to be where it was going to be the best. But now it turns out that I have a concern.
I listen to this podcast and I'm like, maybe I should have it available in China. Maybe people want to know about my thing.
And so I've got some web forms that post back to got, I got a bunch of videos on a third party hosting platform.
I've got an asset management system that's also located on shore here in North America. I've got a product information management system that feeds me images that I can't move overseas. So I'm going to have to keep, I can't practically move it overseas. Let's just.
But, and my videos don't play, let's say they're on, I don't know, Vimeo or YouTube or something like that. They're not going to play. They're just not going to play. So how would, how would an engagement or how would your solution work for somebody? In my unfortunate circumstance, possession.
[00:21:49] Speaker B: So I would break this down into two things again because we'd have to.
[00:21:55] Speaker A: Be, I gave you a complicated disclaimers out there.
[00:21:57] Speaker B: Complicated use. Not at all. It's more that there's a business requirement and then the web performance requirement. So me speaking to the solutions architect, what I can address from our standpoint is the performance solution perspective.
[00:22:09] Speaker A: Okay.
[00:22:09] Speaker B: What I can't answer for them is obviously whether they have legal requirements that are specific to their industry, especially when they're collecting personal information from visitors in China, whether that be stored onshore in China, things like that. I can't speak to that. I can kind of allude to what they might have to consider. But let's just speak about the performance aspect first. So because everything is built and essentially in North America for the most part, it makes sense, right? Because you are also working with tools that are north american based for the most part, or maybe european based. And even the distances from, let's just say us to Europe, that latency is negligible for the most part. Right. You probably have a number of different tools, but you don't actually look into the resource waterfall to say, hey, this is really shocking. We need to move things closer to Europe, for example. So in the case like this, let's just say they are able to keep, and business wise meant to keep everything in North America. And you just want that global website to work for visitors in China. The first thing that we'd address is that you do not need to move your origin server. Now this is again for the performance requirement. So you would keep your origin server and the technical stack that you've already built more or less where it is.
[00:23:23] Speaker A: Okay.
[00:23:23] Speaker B: What Chinafi does in terms of bolting it on is we kind of sit at a layer parallel. I'm trying to visualize this.
[00:23:31] Speaker A: Okay, good.
[00:23:34] Speaker B: We all don't have diagrams. This is a call.
[00:23:37] Speaker A: Okay, just so everybody knows. Yes, we're going to do a blog post, we're going to do a bunch of fancy diagrams, and we're going to make it so that's going to follow on this. So we have a link to that.
We're going to have diagrams of plenty.
[00:23:49] Speaker B: Diagrams of plenty. But for the purposes of this podcast, what ends up happening is Chinafi generates a version that's essentially complementary to a global version that's purely for visitors in China. When we say purely for visitors in China, it's because at the DNS or the CDN level, we use geoip based routing, so that if a visitor is coming from China, we send them to the Chinafi optimized version of the site. If not, they just go back to the original. So more or less you're maintaining your existing stack and Chinafi, or whatever China optimized version of your site is kind of sitting complementary to it purely for visitors in China. Now, for the purposes of this conversation, the Chinafi component can be broken down into, again, two parts. There is of course, the primary resources which are optimized with a near China CDN. If you do not obviously have the regulatory requirements to host anything in China, so everything still has to stay outside of China. But what we do is we have a fully managed near China CDN, and we do give our users the option to use their own CDN provider should they wish. Right. So a lot of the times, Akamai actually has quite high performing outside of China CDN. Near China CDn. Sasley unfortunately does not perform as well when compared to their peers with this aspect of their tooling. I'd love what they do otherwise. But essentially the Chinafi component includes what you do for the primary resources. Right? So this is where the CDN comes in play then. Chinafi has this proprietary, what we call third party resource optimization suite. It's a very lengthy way, really, to say that there are portions of the website that the CDN and the hosting providers do not address. And these are things like you mentioned. So Vimeo as an embed on the website, Vimeo itself is blocked as a platform, but the content, most of the time as a business website is not something that's problematic. Right. We're maybe just introducing what the company does, that sort of thing. And so what Chinafi has to do is we then have to identify that there's a Vimeo embed on the site, process it and treat it based on the best practice. So in Vimeo's case it is to reprocess a video and deliver it either through a China compatible video player if the user has an account or it's a non branded mp4 video that we can automatically process. So video is a good example because it's visual queuing that the end user can end visitor or user can see.
And then another example would be, let's just say with web forms it's a little bit more complicated because you are passing information through and storing it as you normally would. So Chinafied actually doesn't touch that, it would simply pass through much like it would a CDN. So dynamic requests would still be fetched from the origin server and go pass through chinafied back to the end visitor. And any responses China just take that path of passing through. So private and no cache elements are respected in that aspect. So based on the cache headers that you guys would have set very well and properly according to best practices, listening to this podcast would, then you can make sure that the cache resources are delivered through in your China CDN and things that are not cacheable, if it's private or no cache, simply pass through. And anything that requires some sort of tweaking, whether it's replacement, removal in some cases, if it's a compatible proxy, things like that, those are all happening without much of your input. The Chinafi system more or less handles that and it's done through a set of if this, then that rules.
[00:27:34] Speaker A: Okay, sweet.
All right, a couple of questions. So it seems that in this case there would be some analysis, some engineer back and forth, some solutioning, some bright ideas and so forth. Because there's going to be cases where, especially for high cache hit ratio pages, where you're going to figure out a bunch of these things like oh whoops, you're using ad disks on the bottom of all your pages. Add this is blocked in China.
A similar functionality for this is available from this tool. So we're going to rewrite it like this. And this is how it's going to get delivered to your user.
It's a highly cacheable page. Let's say it's a product information page that gets updated barely ever once a month or something like that. So you can have a long lifetime cash on it. And then here we go. It's all good. Let's just say for other things, let's just say you have some search driven pages that require active interaction with a search engine or something like that, so they're more highly dynamic. Or let's just say you're running some personalization or something like that. Things that would take a lot more time for a round trip back to Ohio and you wouldn't want to do that. That sounds like something where you would actively be working with a solution partner to work through what the best way is to get a low latency solution for the particular pages that are on the site.
[00:29:08] Speaker B: In the case of search, for the most part, I would say search versus personalization. I'd probably speak to those a little bit separately, because with search, most of the time there is a latency, undoubtedly because we are talking about round trips here, but in the bigger picture, it's more or less negligible, again, for the most part. So we're not talking about taking something from the two second mark to the 1 second mark. That would be an ideal goal because when you start talking about that, it's ten times the price, right?
[00:29:38] Speaker A: Exactly. More about taking from the 62nd mark or the timeout mark back to the five second mark or something like that. Like something where somebody could actually be expected to wait that long, correct, exactly.
[00:29:53] Speaker B: Okay, we're managing expectations here. We know that companies do not have endless reserves for every possible milestone, even though the engineer might very much well want those gains. But when it comes to a practical situation, most of the time the process will go like this. So they're presenting us with the issue. They've already decided legally, business wise, this is what we're comfortable with.
Let's just say for this instance, it's entirely offshore and outside of China. We then kind of evaluate the website and we identify, okay, this is where the website currently is.
Our automated rules make capture 70% to 80% of that website. Right. Because of the sheer number of websites we see, we can standardize to some extent, rules to handle most things. So Vimeo and YouTube are things that we handle automatically. So there's nothing that really needs to be tweaked there. Then we kind of go into, let's just say the 20% to 30% like you mentioned. Oh, we need to make sure that the search works. Does the personalization a b testing work? Because a b testing is also a little bit more complicated because we're looking at, they have different ways about it, but essentially we're looking at different parameters. Our system has to identify that this is personalized for this user. Okay, how is this personalized for this user and what are the conditions that need to be met and ensuring that we pass those instructions through and make sure that the personalization engine is still receiving those responses correctly and delivering the correct response to the end visitor. And so you're absolutely correct, this is where we need a little bit more conversation for the most part, if something is entirely blocked. So let's just say firebase. Firebase is something that in a case where we actually have to tell the team building website, look, you do have to build something custom for China in the case of Firebase, because it's just so integrated, it's blocked pretty much the whole way through, no matter what tool you're using. I think messenger exception to the rules for now. But again, for the most part, Google Firebase is blocked. You need to rebuild something. Okay, it's very rare that that happens, but that happens sometimes. But otherwise, for most websites that we work with, 90% to 95% of the time, they really just give us the domain or the subdomains that they want optimized. Our team essentially handles the optimizations and the rules that take place. We then pass the engineer or the web developer a preview staging URL to identify the look and the feel after we've made those tweaks, most of the time they can't tell the difference because it's unless you're going to inspect Chrome and looking through where the resources are coming from, they can't tell. And then if they're happy with it, we go live, which is just a quick update on the DNS or the CDN to do that geoip based routing that I talked about.
[00:32:45] Speaker A: Okay.
[00:32:45] Speaker B: And so that's kind of, I would say for 90% of folks that work with us, they really don't have to think much about it. That's why they like us. Cost wise and emotional pain wise, it's less costly to use something like tranify and then 10%. We work very closely with their engineers to kind of figure out custom solutions to make that special case work for them.
[00:33:11] Speaker A: Sweet. Okay, what is your typical engagement then look like for an interested customer? And what are the sorts of people that end up involved in that?
Do you usually have just the customer themselves? Do you have the customer plus your solution integrators and agencies and so forth that worked on the site? What is that end to end generally look like?
[00:33:42] Speaker B: That's a great question, actually. So if the website is developed and managed by a third party agency, most of the time our conversations will be with the end client as well as the agency. And the reason for that mostly is actually mostly when it comes to just client management, because they're aware of the systems that are in place more so than the end user. So we can quickly ask them any questions that we have about the system, such know, what DNS are you using? Or what CDN provider do you have the requirement to continue using that CDN provider even in the context of China?
And what cms are you built on? Things like that. Oftentimes the end user might not be aware of those items. So in a, let's just say most cases, we will have a solutions architect or the head of it, whoever's managing the technical services internally. An agency, if an agency is involved as well as most of the time ahead of marketing. And that's mostly from a budget standpoint, because depending on who the budget sits with, it could be with the marketing teams or with the IT teams. That's what we've discovered in the course of the process. So maybe on a call we'll have three to five people, and at some point, if it's a company that has a vendor onboarding security compliance team as well, we'll probably have a chat with them at some point. But yeah, I would say from our regular engagement, the first person we talk to is the person feeling the most pain, the person getting the complaints from the end visitor vagina, or the agency who's getting complained about because their client is unhappy with the website.
[00:35:21] Speaker A: So that's our first website that you made, is correct, yeah. The gear that you provisioned is slow.
Dwayne, why did you make it go so slow?
[00:35:34] Speaker B: Single handedly created the whole powerful. Really powerful.
[00:35:42] Speaker A: Yeah.
Okay.
Sorry.
Were you still in the middle of a thought then?
[00:35:52] Speaker B: No, I was just thinking the best way to get started and have a conversation is to have, whether it's coming through you guys or direct to Chinafi. Have us run a test of your website for free, just kind of to see where it's landing in terms of performance today. And there are a lot of tools out there that are available complimentary. They're not holistic, I would say, on the free side. Right. But there are synthetic tools out there. Web page test is a great choice, even though they tend to get backlogged there. Chinafine has subtools as well. Or just reach out and find out where your website stands today. From there, you can then evaluate your solutions. But if your stakeholders or your team is not even aware of the problem, that's the first step. Once they realize that, it's shocking. And for the most part it is shocking. Then they're incentivized to kind of figure out, okay, does China matter? And if so, what are the options that we have to solve?
[00:36:47] Speaker A: This makes sense now, I guess. Another question.
So let's just say a solution is created. Let's say my hypothetical website, that I didn't tell you what its name is, but my hypothetical website, it goes through Chinafi now and it's great.
Let's just say I made a couple of updates, I changed a couple of endpoints and so forth so that my forms work. I figured out a new non on premise search solution so that my search latency is better. So maybe I did a couple of it, or ops or development related solutions on my end to make sure what I do works well with what you do. And now the solution seems to work pretty well. What then do you usually see as everybody knows these rules because of the fact that nobody knows 100% what is being blocked, rate limited or whatever, and why, and that these kind of things change like it worked before and now it doesn't.
What sort of continuous monitoring do you both either perform or recommend that clients do to stay on top of things from an operational perspective?
[00:38:07] Speaker B: Well, I think performance monitoring tools are a completely different set of solutions. So, I mean, just like it is outside of China, it's very rare that a CMS platform will inherently have their own performance monitoring tools available to the end client. But what we do have that's, I would say, equivalent to the status checks out there, is that we're at least aware of when or if our service is down, right? If for whatever reason servers are down, we would be notified. Or if we're getting a server based issue, let's just say a 503, we're at least notified of the infrastructure level. If something is broken or down, it's very rare that that happens. Our infrastructure is built with AWS. So frankly, if our systems are down, probably half the Internet is down. But if that were to happen, at least we're aware of that situation. The second thing that we do is, depending on the tiers, obviously, of people we work with, oftentimes we will have regular checks, both in person and synthetic, that we provide to our end users. Let's just say on a monthly or quarterly basis, just to do a bit of a sanity check on the website itself most of the time, because before we go live, we already run a bunch of tests. We run time series analysis to check a website across a period of time from different locations, et cetera. Most of the time, unless something substantially changes in terms of the content or the website, or someone flags your website as problematic or something like that happens. It's very rare that all of a sudden that website then goes down or unavailable. But the best way, I'd say the foolproof way, is more or less keep an open dialogue. Performance monitoring tools we obviously will be using internally, but we recommend the client to also have their own, if they have their own at all, or at the very least have an open conversation with us if they're ever having a problem with a visitor in China, to reach out. Because like you say, they oftentimes might not be aware of whether that's a single use case or whether that's across the board. And that's something that we can rule out pretty quickly because we're testing from multiple locations. You'd be surprised how often someone will call the ITE team from a company and say, hey, the China website isn't working. Yell at China fire, whatever it is. Right. And it turns out it's because they have an office VPN, they've been turning it on and off, and then they're walking website, including their own, and then it's a fire problem. So at the end of the day, performance monitoring plus human management is at the base of what we have to do when it comes to China because there are so many elements you can't automate. And that's kind of, I would say, the moat to what we do. But it's also why we understand the issue so well is because there are nuances that you have to cater to in this space.
[00:41:00] Speaker A: Yeah, indeed.
Well, great. Well, I think that's a good place to wrap this for right now because if we go much more into this, we're going to have to dive one more level deep and I'm going to start talking about dashboarding and monitoring and it'll get great. Dwayne, did you have any other questions before we tie this one?
[00:41:28] Speaker C: Would. I wish I would have known this when I was dealing with that customer that was based in Hong Kong.
Things so much easier. Yeah.
[00:41:38] Speaker B: I apologize on the behalf of all marketing and educational materials that we have ever put out that did not read.
[00:41:47] Speaker C: No. Other than that, I don't have any additional questions.
[00:41:51] Speaker A: Well, we'll try to get this out plus our subsequent blog post to maximize our SEO benefits so that the ops folk like us don't have to suffer for long and they know that something can be done about this.
[00:42:03] Speaker B: Absolutely. And you know what, the follow up question that I think many people would probably ask is how about the business metrics that follow after? And actually to keep in mind, because performance is step one. Making that website work is step one. Then you've got all these tools and platforms such as WeChat, Baidu, Waibwal. It's just a whole completely ecosystem of tools out there. Know, maybe have a global, non China equivalent. But I think that's probably something we can hint at the blog post just to kind of highlight, because that's where a lot of our other agency partners do shine, is because we focus on the web performance component. But there's so much more that has to happen afterwards and there are a lot of stakeholders in this.
[00:42:44] Speaker A: Yeah, that makes a ton of sense. Okay, well, thank you so much for spending this time with us, Jody. And we'll have to do a phase two of this at some point, but yeah, we'll follow this up with our big boxes and arrows blog post.
[00:43:00] Speaker B: I had a very lovely and entertaining time say that about every conversation in this.
Very entertaining as well. So I appreciate that from you guys, and I hope the listeners enjoy themselves as well.
[00:43:15] Speaker A: Indeed. All right, good.
[00:43:16] Speaker C: Thank you.
[00:43:18] Speaker B: Thank you and take care. Bye. Don't close.
