Ask HN: How to enter the Chinese web market? | Hacker News Reader

Ask HN: How to enter the Chinese web market?

4 pointsfiremanx5 comments

I'm consulting for a small bootstrapped startup right now that is trying to build a web-based business that serves customers in mainland China. The business model itself is a non-mobile, niche market, but requires access to a few U.S. based web resources which we don't control (none are Facebook or Google related). Has anyone in the community had experience building a product for Chinese consumers like this? Our biggest hurdle appears to be constructing something that can access these U.S. based partner sites with any speed - how do you get around this beyond requiring every consumer have their own VPN?

Any other suggestions for a small company trying to make a model like this work?

View on HN

Yes, I work in China facing these challenges constantly. Non-locally-hosted doesn't make sense. Locally host everything. The issue isn't only availability, it is speed. Host in China. You'll probably need an ICP License for this (actually, for anything online), so make sure your customer gets you one (as they're legally accountable for what they agree to host in China, as they should be, and are, the 'legal person'). Alibaba have VPS solutions very similar to Linode or DO. Get your customer to agree and pay for this, as you are legally not able to.

You mention accessing US sites: Is this front end or back end? Locally hosted servers face the same restrictions accessing outside the mainland as home connections do unless specifically negotiated.

Do have mobile, at least as a responsive level. I have no idea who your customer is, but it sounds like they're in China and are not-so-tech. That means they will likely check-up via whatever's convenient, and that would probably be via mobile.

How niche is it? Can they pay with credit cards (which most people in china and can't)? You'll need a local entity if you want to do payments through Alipay etc. Which can be done with a WFOE if you have aprox. 100k in registered capital.

If you're confident that it'll be successful AND your strategy is fairly stable and you just want you product in China then go for it. But if you ever want to go mainstream there, want to take Chinese investment or anything of that sort that requires you to abide by their regulations just forget about it.

As far as the technical side goes, yea, just proxy their connections if you cannot cache it locally. And you want to use shadowsocks or something else that doesn't get hit by the GFW rate limiting super fast. But you'll still want to have 5 or so exit nodes that you can rotate (or bind togather) in case the throughput goes radically down. And yes if you're a popular product it probably can get you banned in China.

Hi, I am working as a business development executive for a startup software company which develops UX prototyping tool for mobile and web apps in China. Actually, companies like us generally visit the international sites through VPN. It is really inconvenient and we often lost connection very much. The speed is also quite slow. We are suffering from that so much.

Serve the assets from your own server(s). Proxy or cache the third party stuff. If you can't do this (e.g. if you're relying on third parties for OAuth) then consider removing those dependencies, or at least have graceful fall-back if clients fail to load any third party stuff. (You can test this with uMatrix, by blocking all third party connections.)

Create a redundant setup across multiple popular cloud providers (ec2, gce, azure, digitalocean). Replicate each node type in your architecture at each cloud. At any given time, there is a high chance that at least one datacenter will be available. Setup monitoring infrastructure to route to whichever datacenter is performing the best for a given node type at a given time.

Reply on news.ycombinator.com