How He Turned a Chinese Factory Into a Viral Business

3. AI-ify the Factory

Jason agrees that factories are usually a terrible business. They need constant investment in machinery, and suppliers compete by being one cent cheaper than the next.

He doesn’t try to run the factory himself. The owners he partners with are experienced operators; Jason says the paper factory has been running for 17 years and the plastics factory for 35. His job is to make them more efficient. In his words: “I don’t need to run the best factory, I need to run the most efficient factory.”

Most of that comes from AI:

• Quality control: cameras on the production line use AI to spot scratches and color problems faster than people checking by hand.
• Software built in-house: the team uses AI to build its own tools, including the CRM (customer database) and ERP (the system that tracks orders, stock and production), so they fit how the business works.
• Sales and admin: AI helps the team with outreach, client management and communication.
• A company brain: the team works in Lark, ByteDance’s all-in-one workplace app, and has built a tool on top of it that answers any question about what’s happening in the company.
• An AI agent for customers: next year Jason plans to launch “Ducky.” A client texts it an order and it sends the factory a request for a quote, tracks the order and handles reorders.

The result is fewer people doing the same work or better, and lower spending on equipment. Jason adds new capital for machinery and buildings once the basics are fixed. Paking Duck says it saves customers about 15% to 20% on average, partly because its overhead is lower than many US competitors.

Most factories in China aren’t doing this, Jason says. That gap is his edge.

4. Let Content Do the Selling

Most factories find customers through Alibaba and trade fairs such as the Canton Fair, China’s biggest export trade show. Paking Duck does no trade shows, has no marketing team and has never sent a cold email. Customers come through word of mouth and content.

Jason avoids standard factory-floor videos, because plenty of factories already post them. His team explains how well-known brands make their packaging instead, such as how Tiffany keeps its blue the same on every box. The goal is to teach, not just show the production line.

The production is lean. Every post uses an AI clone of Jason’s voice and an AI avatar of him, with background footage sourced online. Human editors still cut every video.

Jason says this brings in 100 to 200 leads a week. Around 71% are requests Paking Duck can’t take on: too small, outside its product range or in regions where it doesn’t manufacture. That overflow is the starting point for his next move (section 8).

5. Do Business in China the Chinese Way

In China, Jason says, business is “1,000%” about the personal relationship. He is more flexible with some smaller clients than with bigger ones because he trusts them.

His advice for a first supplier visit:

• Bring a greeting gift to show respect.
• The seat facing the door is usually for the guest of honor. Let your host guide you to it rather than taking it yourself.
• When you toast, the person holding their glass highest takes the most honor. Don’t raise yours above your host’s straight away.
• Don’t stick your chopsticks upright in the rice.
• If you don’t like a dish, don’t make a big deal of it. Eat what you can.
• Never negotiate at the dinner table. Eat and drink together, then talk business the next day.

The common mistake is arriving with assumed leverage and coming in hot. Mutual respect comes first. Jason also found that being young helped: older suppliers were willing to take a trustworthy young founder under their wing.

For money, he recommends Hong Kong as the gateway. Open a multi-currency account with a Hong Kong bank such as HSBC and pay suppliers from there. Opening a bank account in mainland China is much harder for foreigners.

6. Prove the Market Before You Buy Inventory

Jason has never taken outside investment, and he has come close to bankruptcy twice, most recently just before Paking Duck. The pattern he describes is common: a good year leads to bigger bets, and too much cash gets tied up in stock and new products.

A new SKU (a single product variant) can mean nine months of development plus certification, insurance and legal costs. If it doesn’t work, a self-funded brand can lose a year or two.

So if he started again in ecommerce, he would begin with dropshipping: selling products that the supplier ships straight to the customer, so you never hold stock. It needs the least money and proves demand fastest. The budget that would sit in inventory goes into growth instead.

He points to the founders of Audien Hearing, an affordable hearing aid brand he supplied early on. Jason says they started by dropshipping to test demand, when many hearing aids cost around $2,000. Once the demand was there, they put the profits into engineering and their own patents. Audien now reports more than $100 million in annual business, and Jason says it got there without outside funding.

Dropshipping is a stepping stone, not the end point. Use it to prove the market, then spend money making a better product.

7. Invest in Brands and Software

Jason also invests through Pughaus, his holding company. It now has stakes in 35 to 36 companies.

The deal flow came from building in public. In 2019 and 2020 he posted on Twitter about everything he was testing on his brands, at a time when most founders kept that secret. Software startups wanted brands to try their products, so they gave him free access in exchange for a mention. Those relationships led to advisory roles, where he helped them work out which brands to sell to and how, and to the chance to invest in their early funding rounds.

How he invests:

• Small checks, fast decisions. Most are angel checks (early, personal investments) of $50K to $75K. He meets the founders, in person if he can, and puts the money in if he likes them.
• Back the founder, not the idea. Many of the companies have changed direction since he invested. Good founders pivot when they need to.
• Mostly software. Jason calls ecommerce one of the worst sectors to invest in: survival rates are low and cash is tight. So most of his portfolio is SaaS (subscription software), with consumer brands where he knows the founders and can help, for example with an introduction to Walmart.

Paking Duck has just launched its own fund, DDG Capital (Duck Duck Goose), to invest in its own clients. Packaging orders show how fast a brand is really growing: the size and frequency of its purchase orders. Jason already knows the founders from working with them. Brands take the money because they want a supplier they can depend on, and investment comes with better terms and extra support from his team.

The lesson for founders: your operating business can show you which companies are working before anyone else sees it.

8. Borrow Ideas From Other Industries

Jason doesn’t credit books or courses for his ideas. He watches how other businesses work and asks how the idea would apply to his.

Mercedes often launches a new model with an Edition 1: a limited first run with its own styling that collectors want. Jason copied the idea at Doe: the first 500 units of every new SKU came in different packaging, and his biggest fans bought and collected them.

His next plan for Paking Duck comes from hotels. Marriott and Hyatt own very few of their hotels, yet guests know what a five-star stay will feel like in any country. Factories have no equivalent. Jason wants to create a standard for packaging factories based on machinery, people, trading history, output, certifications and working conditions.

Factories that meet the standard would join a network, and Paking Duck would send them the leads it can’t serve itself. Software and AI agents for those factories would follow. The first step is to apply the standard to Paking Duck’s own factories.

9. Get Yourself Out of the Trenches

If he started over, Jason’s first hire would be operations. It takes so much time and attention that a founder stuck in it has no room left for the creative thinking only they can do.

His other habits:

• Calls are for discussion, not reports. If it could be an email or a brief, send it before the meeting. If a 30-minute slot is done in 11, end it at 11.
• Don’t micromanage. He calls it the most overrated founder habit.
• Keep talking to your team. At more than 45 people, Jason sets up one-to-ones and group calls so new hires know they can still reach him.
• Use one shared workspace. Lark translates automatically, so Chinese- and English-speaking staff work in the same chats and video calls.

Put It to Work

Write down the request your customers bring you most often, and the supplier relationship you trust most. Then ask what you could add that the supplier can’t add alone: demand, better processes or AI.

Build the relationship before you ask for anything. Prove the model with one partner, let content bring the leads, and use the result to win the next deal.