Why do 90% of web architects ignore "The China Problem?" [TRANSCRIPT]

This is the transcript of a conversation between Marta Cukierman (StreamX) and Tad Reeves (Arbory Digital). It’s been lightly edited for clarity and includes timestamps to make it easier to follow.

Still, the best way to experience the discussion is to watch the full recording at the link below.

Tad Reeves: Welcome to Arbory Digital Experiences. This is a special episode that we're doing today with me, Tad Reeves, Principal Architect at Arbory Digital, and Marta Cukierman, who is the co-founder of StreamX, which is a digital experience mesh. She's joining me from Poland and I'm here in the United States. And we're going to talk about web performance in China.

Marta Cukierman: Hello everyone, thanks for having me. 

Tad Reeves: This is going to be a really interesting combo because we're going to be presenting — let’s see, I’m going to be presenting with one of your principal engineers, Kamil Chociej  — yes, in less than, gosh, what is it, like two weeks away or something like that? Two and a half weeks away or something like that. I'm not nervous.

Marta Cukierman: [laughs] Why would you be!

Tad Reeves: Yeah! So we're going to be presenting at  adaptTo() in Berlin. And we're going to be presenting solutions for web performance in China.

But we wanted to give a little bit more characterization than we can get done in that presentation, which is going to be a technical presentation to a technical audience. We wanted to give more characterization as to why this is a problem, what kind of problem we've seen it be for companies and brands, and why we even got into this in the first place. So there we go.

Marta Cukierman: I remember when Tad came over to our booth, well it would be a year and a half ago at Adobe Summit. That's right, 2024, right? And you, while discussing the solution, came over and said, “Oh this should be a good solution for the China case.” And we were like, “What China case? What is that?”

We dug into that later on, but I'm super curious to learn what made you interested in that from the first place.

Tad Reeves: Okay, little bit of story time here. So in addition to being an engineer, I had a brief attempt at an e-commerce business. We were manufacturing things in China and bringing them into the United States and selling them on e-commerce platforms like Amazon.

So I had a couple trips to China as part of that. I had already been an AEM engineer for some time, and I was very familiar with China’s Firewall. And I always assumed this was limited to some things you couldn’t get to — I thought it would just be like sites being blocked. I’d arrive in China and there would be some sites I couldn’t get to, and the rest of everything would be just fine, or at least the same as everywhere else.. 

And I was entirely incorrect. All my expectations were wrong. It really shocked me how much of a different landscape it is. It’s like two separate internets operating with a loose connection between them.

The variability in performance and availability of normal sites — I was trying controversial sites like Facebook or Twitter, and I thought, “Okay, these will surely be blocked.” But it wasn’t just “this site is blocked.” Sometimes they’d let you through once, then progressively block it. They do rate limiting, progressive blocking, slowdowns. Sometimes it’s outright blocked, but often it’s different.

And for normal sites — not controversial, just browsing for a tractor or looking at a watch — sometimes those are blocked or ridiculously slow too. That’s where I got interested. This isn’t just political. This is a global problem that every brand faces. That’s what made it interesting to me.

Marta Cukierman: And I think that also goes back to what the Great Firewall really is. When we say it, it sounds like one system, but in reality it’s a patchwork of technologies, regulations, infrastructure layers. Together they control and monitor cross-border traffic.

Tad Reeves: Correct.

Marta Cukierman: So the anatomy, which is of course unknown in its entirety, is pretty complex.

Tad Reeves: So when you started looking into this, because we both started doing independent technical digging, what did you see brands doing about it — both AEM and otherwise — in terms of China?

Marta Cukierman: My first visits were more driven by curiosity than proper research. And me being me, I like my jewelry, I like my fashion things. So I went there expecting that maybe some of the rich assets would be blocked or throttled.

I’d look at a jewelry product detail page from the EU — it would give you “what else is in the set,” “what other people viewed,” things to add to your order. Then I’d go into China — not the same site but the same brand site — and for the same product it would just be an image, short info, and that’s it. Very little experience.

It got me thinking: to get that extra stuff you need the request calls. And those calls go to the origin. Building that experience from all those little bits and pieces is harder when communication isn’t simple, and when you don’t know whether you’ll get a response or not.

Another thing: sometimes the deeper you go it actually gets slightly better because there’s less stuff to go through. But for one fashion brand I liked, the Chinese homepage was just one image and two paragraphs of text. Right now they’re not even working at all — I don’t know what happened in the last three or four weeks.

But when you go to their global site, you have everything — little movies, carousels with clothing. You think, “Oh, I like this, I like that.” And then on the Chinese site it’s just one picture, two paragraphs, and you think, “Is it broken? Something’s not right”.

Tad Reeves: From a technical perspective, there are some things we know, why would that be. A lot of global cloud vendors don’t operate in China. If your site uses those — whether you’re on AEM, Shopify, WordPress, whatever — if you use cloud services for search, relevant products, product information management, commerce, personalization… those services simply aren’t available in China.

So brands face a crossroads: Do I make a brand new site? A terrible stripped-down site? Try to recreate all functionality? What’s the right thing to do? It puts them on tough technical crossroads.

Marta Cukierman: I agree. And often that happens at the very end of the project. You think of China as just another market, maybe you’ll localize campaigns or images, but you don’t think you’ll need a separate stack. You don’t expect things that work everywhere else to fail.

Tad Reeves: Exactly. And that’s where the problem is. Most people, whether product managers, technical folks, or creatives, simply don’t know there is a problem.

Tad Reeves: You and I both have been in the business long enough to know what it was like before mobile proliferated, and before everybody had a mobile. I was doing web development, and I was a website owner for a major brand. The iPhone had just come out, and I thought, “What does my website even look like on an iPhone?” I had no idea. I didn’t have one.

So I pull it up on an iPhone and it’s completely unusable. My thought was, “Nobody has an iPhone, I don’t need to deal with this right now.” Later of course, analytics showed mobile traffic was growing and suddenly it was a problem I had to solve.

I feel like this is the same. Brands are like, “I don’t have an iPhone so my site is fine.” In this case, “I don’t see China traffic so my site is fine.”

Marta Cukierman: Exactly. That’s fine if you operate local. But if you’re a global brand and you ignore 20% of your potential market — that’s not neccesarily a good approach. You said it's a kind of a visibility problem. Did you encounter situations where people couldn't see the problem?

Tad Reeves: There are various degrees of being aware that this is a problem. Something that hit me hard on my second visit to China was seeing more of the country. If you've been to Europe or the United States, nothing prepares you for the physical scale of Chinese cities. They are so big.

In the U.S., part of our problem is being isolated from the rest of the world. I don't know how much this is a problem in Europe, where the culture is so much more interwoven, but in the U.S., we just have this "over there-ism.

We have an idea of China from the 1950s, with people in huts tending to rice paddies. It's a silly, outdated view. You go there and you see the size of it. For example, I was dealing with a factory in the city of Fuzhou. I could ask a thousand people in the U.S. where Fuzhou is, and I guarantee you almost every single one of them doesn't know. Fuzhou is a city that's larger than Berlin and Paris; it's on par with Chicago in terms of size. It's like Chicago and Los Angeles combined in terms of population. It’s massive. We’re talking about four to six million people in the city itself, and nobody knows it.

And that's just one of tons of these cities all over the place. Never mind Guangzhou and Shenzhen, a metropolis with more people than the entire country of France. It is mind-boggling.

These are people who all have smartphones; they were driving around in Lexuses. While there is a lower class, there are also lots of middle and upper-class people—so many customers who would buy people's things. I went to factories and saw young workers, in their twenties, with their AirPods in and their smartphones, listening to music while they were making things. They would obviously be customers of any global brand if the brand had a presence. This is why I think the idea that "China has all these people, but they're not my customers" is a comforting lie people tell themselves to keep ignoring the problem.

Marta Cukierman: And if you look at your traffic analytics, you may continue to say that.

Tad Reeves: There are reasons for that. This is something I'm going to get into in detail in our presentation. We did a survey where we asked AEM architects: “In your project, do you know how your site performs for users in China?” The result: 89.3% said no. Only just over 10% said they track their China performance. The rest either spot checks it or, actually a statistically significant percentage said they'd "prefer not to know."

Marta Cukierman: And you have to want to know, to know. Because if you're checking from your machine in Frankfurt or Atlanta, it works.The endpoints are on your side of the Great Firewall, so you don't have to cross it. Your audience, however, does have to cross it to access the endpoints you're calling. I've been told "it works on my machine" more than once.

That's another thing that's mind-blowing—you have to want to know, or you won't.

Tad Reeves: The problem is that many people have the idea that it's a regular firewall. Regular firewalls have a set of rules that are always followed. If Reddit is blocked on your corporate firewall, it will be blocked every single time. In China, it's so much more variable. It's based on your location and whether you're in a hotel frequented by Western guests.

I experienced this myself. I stayed at two different hotels in Guangzhou. At one, which hosted people primarily attending the Canton Fair, I could access everything: Google, YouTube, Twitter, and Facebook. Nothing was blocked. Then I went across the street to a different hotel that was still nice but had mostly Chinese visitors. I couldn't get to anything. Gmail, YouTube, and Facebook were blocked. Sometimes a page would load one image and then stop, or it would load the first time, and then refresh and stop. It was a super variable experience.

Then I tried testing my site from a data center. I thought, "I can't get to my own website, let's see if a professional tool works." I used WebPageTest, which has servers in a data center to run browser sessions. It worked perfectly on WebPageTest. That's when I realized the problem.

If I'm a site owner and I test my site's performance in Beijing using WebPageTest, I'll see it looks good. I'll think, "We don't need to fix anything, guys." I won't understand why we're not getting any traffic, and I might assume it's because people in China aren't interested in our site.

Marta Cukierman: And it’s a vicious cycle. On one hand you don’t get traffic because you’re blocked. You can’t see that because you use the wrong tools. Then you conclude maybe there’s no interest.

Tad Reeves: And that's probably not the case, because the problem is even more multi-layered. You run into issues like search engine indexing. Say you made a Chinese language page, which should theoretically be picked up. 

Google isn't a factor in China; Baidu is their 800-pound gorilla search engine. In the rest of the world, you submit a link and Google indexes it, but Baidu won't index your site unless you submit a link, and you can't submit a link unless you have a Chinese phone number, a Chinese passport, or a business license in China. You simply aren't getting indexed because you aren't physically present. So, claiming, "People in China aren't interested in my site," is completely false. It's the wrong reason; the door isn't even open for them to find your site.

This leads to a market standard where people feel it's okay because the problem is so complex. It's like in 2008 when redesigning your entire website for an iPhone was a huge, expensive problem. 

Marta Cukierman: From a brand perspective, you need a partner to tell you there are options. Sometimes, we rely on this accepted market standard: "It just doesn't really work that well over there because it's complicated." The existing ways to make it work seem expensive, so if China isn't your primary market, people decide not to incur those expenses. That standard needs to change.

Tad: I agree. I think a lot of these are just excuses—little white lies you tell yourself to avoid work. I remember similar excuses, like reading articles where people claimed, "I'm not going to change my booking website for mobile because nobody's going to book a hotel on mobile." Of course, these days, nobody could get away with a phrase like that.

 Marta Cukierman: They'd call you out on that immediately!

Tad: “Who's going to shop for and purchase new clothes on mobile?” Ask my 16-year-old daughter. I'm sure she doesn't even know what the Brandy Melville website looks like on a desktop. 

As soon as she got a job and could make money, packaging started arriving—all ordered on her mobile. So, while that perception has obviously changed for mobile, regarding the China market standard, there are still a lot of excuses.

Tad Reeves: What do you think it will take to change that perception of what the market standard should be?

Marta Cukierman: First, it’s important to say not everyone has an awful digital experience in China. There are brands that invest in a local presence, with physical boutiques or supermarkets, and they put money into building a separate website. When you dissect why they do that and whether it's necessary, you find they were given no other choice.

They believe they need that very separate system so that none of the requests need to cross the Great Firewall. They're working in an isolated market with an isolated solution, but that gives them the reliability of the experience. It is possible. Now, I think what's important is seeing how we can make it more affordable.

Tad Reeves: I think that's part of the defeatism people feel—that there's nothing that can be done, or that they can't solve it for an appreciable amount of money they could ever get approved.

I don't think it's true that every website needs to perform well in China. You have to ask yourself: if I waved a magic wand and gave you 100 million high-performing user sessions from real Chinese users this month, would that change your business?

Marta Cukierman: Well, for us, maybe not so much, but if I had a brand selling handmade jewelry that I could ship to China, then yes, absolutely!

Tad Reeves: For example, if your business is a tax preparer for the U.S., I could give you 100 million user sessions in China. Maybe a couple of those are U.S. visitors...

Marta Cukierman: ...on vacation in China, who realize, "Oh look, a tax preparer! I forgot to do my taxes." But that would be very limited, and you probably wouldn't put money into it. Literally, none of those people are going to transact in a meaningful way with your business.

So, while there's that percentage of people who listen and think, "That's a problem other people have, too bad for them," for so many others, there are universal goods. If you're selling cars, you want to sell them in Europe, America, and China. Why not maintain well-performing global businesses in that part of the globe? But you have to make your offer visible and enjoyable to interact with. When the experience is painful or feels broken, that's not how you build loyalty or customers.

Tad Reeves: I also want to mention another angle, which Jodie from China shared with me. Sometimes, you only care about a handful of really important visitors in China; it’s not about the masses. It could be an investor, or someone who wants to potentially hire your consultancy services for a project. They want to know if you're for real. If your site is broken and doesn't load, that one really important visitor gets turned away.

I've known for a long time that when you look at aggregate web stats, you forget that sometimes success comes down to one person having a good user experience. That one person, effectively communicated to, might feed your team for months. I experienced this myself: I put effort into a blog post and presentation, posted it on Reddit, and got five upvotes—one of which was mine. I thought it had flopped, but one of those upvotes was from someone who saw the presentation and engaged the company.

So, when people say, "Only 6% of my traffic is coming from mainland China, I'm not going to invest," you might want to reconsider: Are any of those people important to you? Could they make a huge difference?

Marta: That's one end of things. The other is what I mentioned: brands with a physical presence are investing huge amounts of money into those physical stores. Then you try to access their website, and it's broken. For example, the only version of their site with Chinese might be from Macau, which is on the other side of the Firewall. Or the global site, which works okay-ish, defaults to French, and you can't even figure out how to change the language.

People don't come to the store before they've had an almost intimate digital relationship with the item—looking at it from different angles, learning about it. If that part is broken, you have a very poor start.

Tad Reeves: So, let's say a brand has accepted this point and is now willing to consider China. What steps should they take to approach this? What would be your first thought?

Marta Cukierman: My first idea, my diagnosis from all the observations I've made, is that to effectively and reliably serve your audiences in China, the tech stack has to be entirely accessible without crossing the Great Firewall. That's the foundation. How you achieve that is a different story, but that is the foolproof way.

That does require you to obtain certain licenses, like the ICP (Internet Content Provider) license, so you have to put in a little more effort. But if you are seriously interested in those audiences and that market, and you're making other investments there to gain customers, this is something you must consider. It feels like people do steps one, two, three, and four, and then they tackle the most important one halfway through.

Tad Reeves: I agree. I think for me, Step A is even before getting technical. I love talking networking and gear because that’s who I am, but the precursor is vision. I want somebody to pull up 4K video streetscapes of Shanghai, Guangzhou, Shenzhen, and Shenyang. Look at these massive cities—the largest city in the world isn't Tokyo; it's a city in the center of China nobody's heard of, with 20 to 40 million people, depending on how you measure it. It looks like a sci-fi movie with monorails, skyscrapers, and LEDs everywhere.

Go look at that and see if you can visualize your customers in those streetscapes. These are technically savvy people with means, and all of them have access to the internet. If you can't visualize your customer there—like if you're a U.S. tax preparer—then great, we stop before you get too deep.

But after that, then I think what you said is astute: you have to measure what your customers are getting right now. And it's going to really shock you. It's really bad; it's worse than you think. The one conclusion almost every person comes to is that without a physical presence—where your gear is physically located inside China—the performance is going to be unacceptable. You and I have done a bunch of testing on this, and it's always unacceptable.

Marta Cukierman: It's worth saying that the worst part of the experience is that it’s so unreliable. One day it's okay-ish, and the next day it doesn't load at all. You think, "I visited it two days ago and it kind of worked, and now I can't even get through two or three clicks." That's what makes it so difficult.

Tad Reeves: Understanding that variability, and defining "bad," is key. It's not that a site that usually takes two seconds to load in the US now takes six seconds in China; a user will still browse that. We're talking about 120 seconds to timeout with broken images and buttons because the JavaScript didn't load—that is the actual experience they're getting.

Sites I was proud of in the U.S. just fail to load in China. They take two minutes to fail. I used to think, "It can't be that bad for everyone," so I threw Real User Monitoring (RUM) on it. I had to get RUM hosted in China because tools like New Relic don't work there (they explicitly state their services don't work in China). I got an alternative and stuck it on my site. I told friends, "Go browse this and tell me what you get." The real-life stats showed, "Your site loaded. It took 1 minute and 9 seconds."

Marta Cukierman: Okay, then there's a question—and I think I already know the answer—would you say that people over there are just used to this kind of performance and are okay with it?

Tad Reeves: That's the thing: there is no shortage of local competition. There is so much local competition for everything, and it loads quickly because those sites are hosted in China. If you're a global brand selling something, there is almost certainly a local brand whose site will perform.

Previously, a global brand might have had a small excuse because some Chinese consumers were interested in foreign goods and would wait. But now, you're already fighting at a disadvantage, especially with patriotic consumers who prefer to buy Chinese anyway. If your site is terrible, I don't think they're making any viable excuses for you.

Marta Cukierman: I don't think so either. They're wired not to wait. Like every other internet user, we hate waiting. It's not acceptable user experience. They're not accepting it, just as we wouldn't be loyal to a brand that makes us wait; we'll search elsewhere.

Tad Reeves: Oh, yeah. Even if you're a captive audience. I have a travel booking provider I've used for 15 or 20 years, and their mobile performance is not good. I just had an experience waiting almost three minutes at a hotel desk for my reservation to load. I was thinking, "I'm going to switch from these guys; this is the last straw." That's the same experience my own website gets in China, and they're not a captive audience.

Tad Reeves: Anyway, the positive end of all this—because we need to stop any brand owners from committing self-harm as a result of this podcast—is that the worst belief is: "There's not something practical I can do to solve this problem." 

This came through in our survey results. People believe that if there was an easy, practical solution, they'd use it. But when you look at big vendors like Adobe Experience Manager, people think there's no practical solution.

We do have a solution that works pretty well right now. It's neat, right?

Marta Cukierman: It's very neat and flexible. It's not just tied to one specific variant of the Adobe stack; it can be adapted effectively and cost-effectively for a number of different solutions.

I don't want to turn this into a sales pitch, but the main thing people need to know is that solutions actually exist. It's not the direction you should take to just say, "I'm pretty sure I can't do anything about this right now."

Tad Reeves: You can absolutely challenge that assumption. Now that you know there are ways to do it, go find out what it is! If you really want to know, watch Tad's and Kamil's presentation at adaptTo(). You still have a couple of weeks to wait, but be prepared to be amazed. It's working, and it's solving this for different setups. This is going to be a fun new era. I made a decision back around 2018 that I really wanted to help global brands get their sites to work in China, and I feel like this is one of the first times where it's truly possible. 

I've encountered roadblocks in the past, like the requirement for a local office in China, but that is also surmountable. There are solutions for almost every business configuration. If you have the desire and an audience you want to reach, you can actually solve this problem.

Marta Cukierman: Exactly. The technology has caught up, and you can now expect a different standard.

Tad: Yes. It will be interesting to do this podcast again in a couple of years when we've turned the tide on this, and a lot of global brands are hosting there and achieving good results. It would be great to have a different set of expectations for China.

Marta Cukierman: Yes, I say a big yes to that.

Tad: Okay, well, I think that wraps us up for now. The next time we meet, I'll be on the same continent! We already have plans for a wrap-up after  adaptTo(). Also, because I was asked, we're going to be doing a pre- adaptTo() bike ride and video. For anyone who can make it to  adaptTo() before the conference starts, we're doing it on Sunday. We'll do a little bike tour of Berlin and get people pumped up so they can...

Marta Cukierman:  That's right, absorb all the great stuff that will be on stage later over the next three days!  adaptTo() is so much more technical than Adobe Summit; you really have to be ready to receive that information.

Tad: Yes, there's a lot of meat there, very little fluff. And we aim to keep our presentation that way, too. 

Cool. All right, thanks, Marta, for coming on.

Marta: Thank you for having me!