Consumer Brands Run Like Internet Companies
Ship physical goods like software: small-batch launches, then scale only the winners.
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 8
- Confidence
- 72%
Instead of betting on a single hero product, the brand is operated as a software team operates a product: launch constantly, measure ruthlessly, scale only survivors. Contract manufacturers in China evolved from build-to-spec (OEM) to design-and-build (ODM) to near-complete brand production (OBM), which collapses the cost and lead time of a new SKU. A brand exploits that by releasing new products on a weekly or faster cadence in small batches, testing them simultaneously as physical goods and as marketing creative. A small share, roughly one in twenty by the founders' own accounts, clears the bar and earns placement in expensive mass offline channels; the rest are killed. Because offline distribution is the genuinely scarce resource, the discipline is not in inventing products but in refusing to fund the ones the data did not select.
Origin
Described by Rui Ma and Ying-Ying Lu on Acquired as the pattern behind a wave of Chinese consumer brands including Genki Forest, Perfect Diary and Shein, which import internet-company operating habits into physical consumer goods.
Core principles
- 01A flexible contract supply chain is a testing instrument, not just a cost centre.
- 02Product launches are experiments; most are supposed to fail.
- 03Distribution is scarce, so only proven SKUs deserve it.
- 04Marketing ROI discipline transfers directly from gaming and internet teams to physical goods.
- 05Own the brand and the demand data; rent the factory until the brand is proven.
How to run it
- 1
Pick a category with foreign-brand overhang
Target consumer categories where imported brands hold a share far above what a developed domestic economy would predict. That gap is the market you are attacking, not a nationalism story.
Pro tip Benchmark foreign-brand penetration against a mature market to size the gap before committing.
- 2
Rent the supply chain instead of building it
Contract with an ODM or OBM partner that can design, manufacture, and even draft marketing copy. Your scarce capital goes into brand and demand, not plant.
Watch out Supplier claims about full-service brand capability are partly marketing; verify what they actually design versus assemble.
- 3
Staff the team from internet operators
Recruit leaders who think in marketing ROI, A/B tests and iteration cycles rather than classic brand management. The operating cadence follows the people you hire.
Pro tip Gaming and performance-marketing backgrounds transfer unusually well because both live on live-ops experimentation.
- 4
Set a launch cadence, not a launch date
Commit to releasing new SKUs on a fixed rhythm, weekly or faster, rather than developing one flagship product. The cadence is the mechanism; individual products are disposable.
Pro tip Collaborations and seasonal variants are a cheap way to sustain the cadence without new tooling.
- 5
Test in small batches on both axes
Run each new SKU as a limited physical batch and as a marketing creative test at the same time. You are measuring demand and message together, not sequentially.
Watch out Small-batch economics only work where minimum order quantities are genuinely low; check MOQ before designing the process around it.
- 6
Apply a hard survivor threshold
Set in advance the share of launches that earn further investment and kill the rest without debate. In practice only a small minority clears the bar.
Pro tip Write the threshold down before results arrive so winners are chosen by data rather than by advocacy.
- 7
Spend distribution only on winners
Mass offline channels are hard to penetrate and harder to manage, so reserve them for products the tests already selected. Everything else stays online.
Watch out Pushing unproven SKUs into offline retail consumes the working capital the testing loop depends on.
- 8
Vertically integrate late, if at all
Bring manufacturing in-house only once the brand and the winning SKUs are established, as several of these companies did years after launch and close to listing.
Pro tip Treat your own factory as a margin decision, not a credibility decision.
In the wild
Genki Forest launched into carbonated beverages with a team drawn from social gaming, so the company reasoned in marketing ROI and rapid A/B tests rather than classic beverage development. New products appeared every few days in small batches, tested as physical goods and as marketing at once. Only about five percent were pushed into mass offline channels, which are expensive to win and to manage. The company did not open its own factory until years in.
→ A roughly six billion dollar valuation while explicitly targeting the Coca-Cola position in China.
The cosmetics brand behind Yatsen ran a similar loop, leaning on collaborations and seasonal products to sustain roughly one new product per week. Whatever became a bestseller received more resources; the rest were dropped. Like Genki Forest, the company only stood up its own factory shortly before going public, having used contract manufacturers to keep the iteration cost low.
→ Listed publicly while positioning itself as the L'Oreal of China.
Shein pushed the same logic to its limit in apparel, producing around a thousand new designs a day at an average of five days from design to shipment, selling exclusively outside China. Demand was driven almost entirely through social media, with roughly nineteen million Instagram followers at the time of recording, and prices low enough to make each design a cheap test.
→ Around ten billion dollars in revenue, and the benchmark almost every cross-border Chinese fashion startup now pitches against.
Common mistakes
Building the factory before the brand
Owning manufacturing early locks capital into fixed capacity and destroys the cheap-iteration advantage. The brands that worked contracted production first and integrated only years later.
Believing the supplier's brand claims
Factories now market themselves as full original brand manufacturers, but much of that is positioning over what are still design-and-manufacture shops. Verify what a partner actually does before outsourcing brand-critical work.
Running the cadence with no kill rule
The loop only creates value if most launches are discarded. Without a pre-set survivor threshold, weak SKUs absorb the scarce offline distribution the winners needed.
Is it for you?
Best for
Founders in fast-moving consumer categories with access to flexible contract manufacturing and cheap online distribution.
Not ideal for
Regulated, capital-heavy, or long-lead-time products where each iteration takes months and small batches are impossible.
From the transcript
“And that is basically how if you think about it, it's very similar to how you iterate on an internet software product, right?”
“according to their founder, only 5% of their products make it into sort of mass channels, right? Because the mass the mass offline channels are…”
“I think their average is one new product a week, and then whatever makes it, you know, quote unquote becomes a bestseller, then they'll put…”
From the episode
Special: 2021 China Tech Trends (with Tech Buzz China)