Frequently Asked Questions

Great Firewall. China runs a national filtering system and it blocks or throttles most of the Western services that WordPress sites tend to depend on. We’re talking Google Fonts, Google Analytics, reCAPTCHA, YouTube, Facebook pixels, most of the major CDN providers. All blocked or so slow they might as well be. We’ve seen sites that load in under two seconds from New York take 15 or even 20 seconds from Shanghai. Sometimes they just give up and timeout. If nobody has done any China-specific work on your WordPress setup, your site is basically invisible to anyone in mainland China. That’s not an exaggeration, it’s just what happens.

Yeah you need one, and there’s no clever workaround. ICP is Internet Content Provider. It’s a permit issued by China’s Ministry of Industry and Information Technology and the law requires it for any website hosted on a server in mainland China. There’s two types: ICP Bei’an (备案) which is for informational websites, and ICP Jing Ying (经营许可证) for sites that process transactions. Without the Bei’an, no Chinese hosting company will serve your domain. They just won’t do it. We manage the whole filing. Usually takes two to four weeks, sometimes a bit more depending on which province the server is registered in.

Four to eight weeks is the range for most projects, though every one is a bit different. Week one is the audit where we figure out what’s broken and what needs to happen. Weeks two through four we deal with the ICP filing which is mostly waiting. Server setup and the migration itself happen during weeks three to five. Plugin swaps and performance tuning fill weeks four through six. Testing and launch usually wraps around week six to eight. If you need the commercial ICP license rather than the standard Bei’an, add a couple weeks. We can usually overlap some of these phases but the ICP wait is the ICP wait, we can’t really speed that part up.

Our project managers speak English and Chinese, and a couple of them speak French too if that’s useful. We overlap with European hours in the morning and pick up North American teams in our evening. Sprint cycles with weekly updates in your inbox, that’s the baseline for every project. Tools are whatever you already use. Slack, Teams, Zoom, email. We don’t really care. We’ve managed projects with teams in Paris and Chicago and Sydney all going at once and timezone has never actually been the thing that caused problems. Usually if something goes wrong it’s a scope issue or a content delay on the client side, not the time difference.

Basically anything that calls a Google API. So Google Fonts, Google Maps, reCAPTCHA, Analytics, all of those. Same deal with Facebook SDK, Twitter embeds, anything sitting on a CDN that the firewall blocks. Replacements we use: Baidu Tongji or CNZZ for analytics, Baidu Maps instead of Google Maps, Tencent Captcha for bot protection, self-hosted fonts for typography, and Alibaba Cloud CDN or Tencent CDN for delivery. Before we start any migration we audit every single plugin on your site. You get a report that goes through each one. Stays, goes, or gets replaced, and with what.

That’s what most clients end up doing actually. You keep one instance on your current host for global traffic and we build a China-optimised copy on mainland servers. Geo-detection and DNS routing send Chinese visitors to the local version, everyone else goes to your global site, nothing changes for them. No need to uproot your whole infrastructure. The two instances can share a CMS backend or be completely separate, it depends on how much your China content differs from the global version. We’ve done it both ways, there’s pros and cons to each that we can walk through.

So it’s split DNS with geographic resolution. Visitors from within China resolve to mainland A-records pointing at your ICP-licensed server. International visitors resolve to your existing infrastructure, totally unchanged. We configure this through DNSPod (Tencent’s DNS product) or Alibaba Cloud DNS, both of which handle geo-routing out of the box. For domains we generally recommend keeping your registrar where it is, outside China, but grabbing a .cn if Baidu SEO matters for you. Baidu gives a small credibility bump to .cn and .com.cn domains. It’s not a massive factor but it helps, especially early on when you don’t have much Baidu authority built up.

The speed improvement is where you see results first. A site that goes from 15 second loads to under three in China, bounce rates drop hard. Usually 40 to 60 percent lower. Conversions follow but the specifics depend a lot on your product, your price point, your audience. The more strategic payoff is Baidu indexing. Once your site is set up correctly for Chinese search you’re reaching organic traffic that literally did not exist for you before. Clients in competitive verticals tend to see solid Chinese search traffic within three to six months. Less competitive niches, faster than that. Hard to give exact numbers without knowing your space but that’s the general trajectory we see.

Our founder moved to mainland China over 20 years ago. Not visiting a few times a year, not advising over video calls from another country. Living and working here, full time, for two decades. Over that stretch the team has delivered dozens of WordPress projects for international companies coming into the Chinese market. Week to week we’re filing ICP applications, sorting out firewall issues, going back and forth with hosting providers in Chinese, submitting things to Baidu Ziyuan. That’s just what the work looks like from our end. If any of that sounds like something you’d want to verify, we’ll connect you with past clients. Happy to do it.

Where to start. Baiduspider, their crawler, is pretty bad at rendering JavaScript. Server-side rendering or pre-rendering isn’t a nice-to-have, it’s mandatory for proper indexation. Baidu still actually uses the meta keywords tag which Google stopped caring about years ago. Sites with ICP licenses hosted on mainland servers get a measurable ranking boost. Sitemaps and indexing requests go through Baidu Ziyuan, completely separate ecosystem from Google Search Console. Backlink value skews toward Chinese-language domains, .cn, .com.cn, that kind of thing. Duplicate content gets penalised harder than on Google. Honestly the whole approach needs to be rebuilt from the ground up. People sometimes ask if they can just adapt their Google SEO strategy and the truthful answer is no, not really. It doesn’t transfer.

Yes, and you should. We hook up both as payment gateways, works fine with WooCommerce or custom checkout setups. Beyond the payment piece we also connect WeChat Official Accounts for social login, set up the Open Graph tags so pages share properly inside WeChat, and link Mini Programs where that makes sense for the business. The thing people outside China sometimes don’t grasp is how central WeChat is. It’s not just a chat app that people also use for payments. It’s genuinely the operating system of daily life for most Chinese consumers. If your site doesn’t play nice with it, you’re losing people before they even consider buying.

Yes. The team is in mainland China. We test on local networks, not through a VPN from somewhere else. We interact with Chinese hosting providers and government portals in Mandarin. We file ICP paperwork ourselves. When something goes sideways at 2 a.m. Beijing time we’re in the same timezone dealing with it, not waking up six hours later to an email chain about the problem.

The reason we emphasise this is because we’ve seen what happens when companies work with agencies that claim China expertise but are actually based in London or Singapore or wherever else. Things work fine until there’s a real issue, and then suddenly the response time isn’t there, the debugging hits a wall because they’re testing from outside the firewall, and the whole thing drags out. We don’t have that problem.

Three layers. Server level is Redis or Memcached for object caching plus Nginx FastCGI cache for full page caching. CDN layer is Alibaba Cloud CDN or Tencent CDN for static assets from Chinese edge nodes. Then browser level with cache-control and expires headers dialed in.

We don’t use WP Super Cache or W3 Total Cache or any of those. They clash with China CDN configurations in ways that are really annoying to troubleshoot. Doing caching at the server level takes more setup work initially but it’s way more predictable once it’s running. We’d rather spend time configuring it right once than chase weird cache invalidation bugs for weeks. Less magic, more control.

Mobile, mobile, mobile. Over 80 percent of web traffic in China is from phones and it colours everything. QR codes are everywhere, sharing happens through WeChat, payments go through Alipay or WeChat Pay. Chinese web design tends to be more information-dense than what Western users are comfortable with. Like, significantly more dense. Different cultural expectation around how much content should be visible at once.

Trust signals are a whole different thing too. Chinese visitors look for the ICP number in the footer, real company registration info, customer service available on WeChat. An email contact form on a separate page doesn’t really communicate “legitimate business” the way it might in the West. If the right trust markers aren’t visible, people leave. It’s not subtle.

Hosting on ICP-licensed servers inside mainland China makes a full block really unlikely. We should be clear about that. Content-level filtering can still get triggered though, that’s a different mechanism. We run monitoring from several Chinese cities using local test tools and because we’re here we catch things quickly. We’ve watched agencies based outside China spend days, actual days, trying to figure out what’s being filtered because their testing runs through a VPN which distorts everything. They’re essentially debugging a different version of the problem than the one their users are actually experiencing. We don’t deal with that because we’re on the same network.

Export with WP-CLI or a direct MySQL dump. Search-and-replace for URL changes, has to be serialisation-safe obviously, we use interconnect/it’s Search Replace DB or WP-CLI’s search-replace. Transfer to the Chinese server goes over SSH/SCP. Post-import we verify integrity on every table. Big databases we sync incrementally so the downtime stays short.

The encoding part is where people get bitten. You need UTF-8mb4 and you need to verify it’s actually set correctly, not just assume it carried over. If encoding goes wrong during migration, Chinese characters break everywhere and fixing it retroactively is genuinely miserable work. We test this multiple times before the cutover. It’s not the glamorous part of a migration but it’s the part that will ruin your week if you skip it.

Straight translation almost never lands the way people expect. Chinese search queries just don’t map from English in a direct way. User intent gets structured differently, the words people use to search for the same concept can be completely different. What we typically suggest: take your core brand messaging and have native Chinese writers adapt it. Not translate. Adapt. There’s a real difference. Then build original content on top of that based on actual Baidu keyword research and local search behaviour.

Running your English blog through translation and posting it to Baidu is not a strategy that works. We’ve inherited sites where the previous agency tried that and the content just sits there unindexed. Localisation is a different skill than translation and the gap between them is bigger than most people think.

Clients from Europe, North America, Asia-Pacific. Industries range from luxury retail and consumer goods to B2B manufacturing, SaaS, professional services, few other verticals mixed in. The common thread is always the same: WordPress plus China, and they needed a team that could actually do both.

We can share relevant case references on a call. Or better, we can connect you directly with past clients so you get the unfiltered take on what working with us is actually like. We’d rather you hear it from them than from us, honestly. Our version is obviously going to be biased.

We deploy on Alibaba Cloud or Tencent Cloud. Both have ICP-compliant data centres in mainland China. Stack is Linux, usually Ubuntu though sometimes CentOS for legacy reasons, Nginx, PHP 8.x with OPcache, MySQL 8.0 or MariaDB, Redis for caching, and a China CDN layer. We pick whichever server region is closest to your audience. Shanghai, Beijing, Shenzhen, there’s a few other options too depending on your use case.

Backups, server hardening, and uptime monitoring are configured before we migrate anything. That’s not an optional add-on that gets discussed later, it’s part of the base setup.

Depends on what you actually need, which I know is an unsatisfying answer but it’s the true one. Migrating a clean existing WordPress site to Chinese hosting is a fundamentally different scope than building a new China presence with content and design and Baidu SEO from scratch. We don’t quote ballparks over email because they’re always wrong and they create weird expectations.

What we do instead: free initial audit, produces a detailed quote broken into phases. Audit, ICP, migration, optimisation, content, maintenance. Monthly hosting and support is separate line items. You see everything before committing. No charges appearing three months in that weren’t in the original scope. If the project grows, we scope the addition separately and you approve it before we start.

Process and visibility. Every project has documented milestones with acceptance criteria spelled out. You get staging access where you click through everything before it goes live. We ship in one-to-two week sprints so you’re reviewing actual work constantly, not hoping for three months that it’ll all come together at the end.

We also maintain a risk log. Sounds boring, very useful. If something is trending off-track we tell you before it’s a problem, and we bring the fix with us, not just the bad news. In our experience the projects that go wrong in this industry go wrong because of communication gaps, not technical failures. So we over-communicate. Some clients find it a bit much at first but everyone ends up preferring it.

Technically same as anywhere else. The piece that’s specific to China is the certificate authority. Not every CA is trusted by all Chinese browsers, and if you pick the wrong one, visitors on QQ Browser or UC Browser get a security warning before they see a single pixel of your site. Not great.

We use DigiCert, GlobalSign, or TrustAsia certs through Alibaba Cloud. Let’s Encrypt is fine most of the time but validation can lag from Chinese networks. HTTPS with HSTS headers, HTTP redirect forced, renewal automated. Pretty standard stuff but the CA selection is the part that trips people up if they’re doing it without China experience.

Yeah, full adaptation. Alipay and WeChat Pay as payment gateways. Chinese logistics integrations, SF Express, ZTO, YTO, replacing whatever Western shipping plugins you currently run. Product pages reworked for Chinese shopping patterns. And critically, a checkout flow that’s tested inside the WeChat in-app browser. Huge number of mobile purchases in China go through that browser and if your checkout breaks inside it you’re losing sales you’ll never even know about because the user just leaves.

We also set up CNY pricing and fapiao support. Fapiao is China’s official tax invoice and it’s required for business purchases. Not a feature, a requirement. If you sell to Chinese companies and they can’t get a fapiao, they can’t expense it, and the deal probably doesn’t happen.

We get this one a lot. Every project has a bilingual PM as the main point of contact, from kickoff through post-launch. Working team includes WordPress developers with five-plus years of China-specific experience, DevOps people who manage Chinese cloud environments every day (not as a side thing), a UX designer who works within Chinese conventions, native Chinese copywriters, a Baidu SEO person, security and compliance staff.

You meet the team that’s going to build it before we start. Same people from the proposal do the work. We’ve heard enough horror stories about agencies where the senior team pitches and then a completely different group of juniors actually delivers. We don’t operate like that and it’s a deliberate choice.

Under the hood it’s DNS poisoning, IP blacklisting, URL keyword filtering, SNI inspection, and deep packet inspection. For WordPress what that looks like in practice: any wp_enqueue_script or wp_enqueue_style that loads from a blocked domain, googleapis.com, gstatic.com, gravatar.com, those kind of things, will hang. Not error out. Hang. The browser just waits, and waits, and your whole page render stalls.

YouTube or Google Maps embeds leave literal holes in the layout where the iframe was supposed to load. AJAX calls to blocked APIs fail silently, so your contact form or search bar or checkout just… doesn’t work, and there’s no error message telling you why.

During migration we trace every single external call. Theme files, plugin code, wp_options table, all of it. Everything either gets replaced with a China-friendly alternative or self-hosted. It’s tedious work but if you skip even one dependency it can break things in ways that are really hard to diagnose later.

Probably, after we fix it. Out of the box? Almost certainly no. The usual culprits: Google Fonts, which is far and away the most common issue. JavaScript libraries loading from blocked CDNs. Unoptimised images with no China CDN delivery. Third-party tracking scripts hardcoded into the theme files.

We do a theme audit and produce a compatibility report. Nine times out of ten we can optimise what you already have rather than rebuilding from scratch. A full redesign only really makes sense if you’re also planning to rethink the user experience for Chinese visitors as part of the same engagement. Otherwise it’s usually overkill.

Those agencies treat China as an add-on service. Maybe they suggest a CDN, or recommend hosting in Hong Kong, and they figure that’s good enough. It’s not, but they don’t know that because they’ve never actually run a WordPress site from behind the firewall long enough to see all the ways it breaks.

We work the other way. We start from China and build out. Active ICP licenses. Testing on real mainland networks. Direct relationships with Alibaba Cloud and Tencent Cloud support. We submit sites to Baidu Ziyuan ourselves. We track regulatory changes that happen way more often than people outside China realise. WordPress in China isn’t a service we added to the menu. It’s the whole menu.

WordPress.org is reachable from China but the connection speed is inconsistent. Some days it’s fine, other days it crawls. We keep local mirrors and run updates through WP-CLI with version control. Every update hits staging first, compatibility checks run, and only then does it go to production.

Updates get scheduled during low-traffic windows. Rollback snapshots saved for every deployment. Some plugins have weird China-specific issues that don’t show up in normal testing, so for those we maintain our own patched versions that we’ve verified on mainland infrastructure. Not exciting work but the kind of thing that prevents a 3 a.m. emergency because an update broke something that only breaks behind the firewall.

Baidu Tongji. It’s Baidu’s analytics platform and it does basically the same things: traffic sources, behaviour flow, conversion tracking, event tracking. The nice bonus is it ties directly into the Baidu SEO ecosystem so your search data and site analytics live in one dashboard instead of two.

For heatmaps and session recordings we install China-compatible tools alongside it. And if GDPR is a concern we can deploy Matomo self-hosted on your Chinese server so the data stays on infrastructure you control and never touches a third party. Whatever setup you go with, analytics is live from day one. We don’t launch sites without it.

WordPress core, theme, and plugin updates, all tested in staging before they hit production. Daily backups stored offsite. Uptime monitoring around the clock from test points inside China. Security scanning and malware cleanup. SSL management. ICP compliance monitoring which people forget about but regulations do change and your license needs to stay current.

Monthly performance reviews where we actually look at the numbers, not just send an automated report. Baidu SEO health checks. Bilingual support for whatever comes up.

Every month you get a report covering uptime, page speed, security, and search visibility. Everything measured from within mainland China. We keep emphasising this point because it genuinely matters. Testing from outside the country gives you a completely different picture and it’s usually an optimistic one. The numbers that matter are the ones your actual visitors experience, and those you can only measure from inside.