Companies Are the Hardest Products, Communities Are the Hardest Companies
Two readings, same week, same room, different worlds — and they converged on the same five principles.
Reading 1: LatePost's interview with Anker founder Steven Yang — a man who lost 15kg to depression in 2022, cut 10 product categories, then rebuilt his company from first principles.
Reading 2: Wang Yizhou's essay People, Content, Community from the Chinese publication Mingyeji — a framework for why most "community products" fail: they solve communication but not identity.
One is about building a $3.6B consumer electronics company. The other is about building online communities that don't die. They look like different universes. They're not.
This post distills the five shared principles — a unified lens I wish I'd had when I started building.
1. Terrain Dictates Strategy, Not the Other Way Around
Anker: Shallow Waters
Yang's most quoted line: "Even with $5-10 billion in chips, don't sit at a table where the blind is $500 million."
Anker categorizes all markets into two terrains:
| Type | Size | Anker's stance |
|---|---|---|
| Shallow sea | ≤ $50B/year | Fights here |
| Deep sea | $100B+ (phones, TVs, PCs) | Won't touch |
This isn't about ambition — it's about what the terrain allows. A $50B category with many sub-niches (like consumer electronics accessories) rewards breadth + systems. A $500B category dominated by mobile phones rewards depth + a single moonshot. Anker chose the terrain that matches its nature: a "third-category company" that wins through portfolio and process, not a single breakthrough product.
REMIX: Cultural Boundaries
The community framework makes the same point. A real community needs five elements:
| Element | Meaning | Counterexample |
|---|---|---|
| Role | A recognized identity | Everyone is just a "user" |
| Action | A specific way to contribute | Everyone can only like/comment |
| Status/incentive | Reputation, influence, money | Pure volunteerism burns out |
| Hierarchy | Levels + sequence + mobility | Flat = no growth feeling |
| Cultural boundary | Anti-scale exclusivity | Anyone can join = no identity |
The last element is the most counterintuitive: a community that scales infinitely is not a community — it's a content platform. Cultural boundaries are the "shallow sea" of community design. If you don't accept that growth will be constrained by identity, you'll build a content platform instead.
The shared principle: The shape of your battlefield determines your strategy — not your vision, not your talent, not your funding. Denying the terrain doesn't change it.
Put it to work: Before choosing a category (product) or a community (people), draw a map of the terrain. What does success at your chosen size look like? If the map shows "deep sea" dynamics, don't enter with a shallow sea strategy.
2. Don't Design for 'Everyone' — Design a Ladder
Anker: The 1357 System
Anker's products aren't a flat list. They're a gradated hierarchy:
| Series | Character | Price vs. peers | Strategic role |
|---|---|---|---|
| 1系 | Cheapest | Lowest | Funnel entry |
| 3系 | Value | Low-mid | Volume base |
| 5系 | Reliable, quality | +50-100% | Core (first 10 years) |
| 7系 | Best, most expensive | Premium | Future |
Target: 7系 grows from 3% (2020) → 15% (2025) → 30% of revenue by 2030.
"How do you judge a world-class company? Look at whether 7系 products are at least 30% of revenue."
The hierarchy gives every user a clear upgrade path without leaving the brand. Solarbank 1 (€1k) → Solarbank 2 Pro (mid) → Solarbank 3 Pro (AI-powered, premium) — same user, same ecosystem, natural upgrade.
REMIX: Hierarchy + Sequence
The community framework's fourth element — hierarchy — isn't about inequality. It's about making growth visible. A community without levels is a chat room. Levels create:
- A path: "I start here, I can get there"
- Recognition: Your effort maps to a visible status
- Retention: Leaving means losing your accumulated standing
The framework calls this "a sequence" — not just tiers, but a progression with clear advancement criteria.
The shared principle: Don't sell one thing. Design a ladder. Let users grow without leaving your ecosystem.
Put it to work: Whether you're building products or a community, map your 1系→3系→5系→7系. Your entry point should be easy to try (low barrier), your core should deliver reliable value (high retention), and your premium tier should set the standard for the entire category (brand elevation).
3. From Factory to Playground: The Remix Model
Anker: The Creator's Paradise
Yang's revised mission for Anker: "Build a creator's paradise that inspires each other. Create a brand that global consumers love."
He contrasts two organization models:
| Model | Used by | Logic |
|---|---|---|
| King + Knights | Apple, Tesla | Single ultra-category, concentric talent |
| President + Federation | Anker | Multiple mid-categories, distributed leadership |
Anker is explicitly the second model. The CEO doesn't design products — he designs the system that produces product designers. The company at scale: 6,000 people, 800+ making >¥1M/year, organized as "branches build, headquarters orchestrates."
Yang's diagnosis after the 2022 crisis: "Five pots, only two lids" — too many product lines, not enough great managers. The fix wasn't hiring more people. It was reducing from 100 products → 30 → ~10 by 2025, then rebuilding the system.
REMIX: The Playground
The community framework's core metaphor: product as playground, not factory.
| Dimension | Factory (traditional platform) | Playground (AI-era community) |
|---|---|---|
| Rhythm | Publish, done | Open every day, ongoing |
| Structure | Platform decides the flow | Platform provides equipment, users decide |
| Goal | User consumes content | User exists in the group |
| Barrier | Low entry, low depth | Low entry, depth optional |
| Repeat behavior | Scroll more | Play deeper |
The critical insight: AI's core capability isn't creation — it's editability/remix. The old product model assumed "seed units" were completed works (tweets, photos, videos). The new model assumes seed units are open and editable — a character, a world, a conversation — and the product's job is to enable remix, not factory output.
Anker's "creator's paradise" is a playground for employees. REMIX's "playground" is a product model for users. Both reject the assembly line.
The shared principle: Stop building factories that output standardized units. Start building playgrounds where participants create value through interaction, remix, and evolution over time.
Put it to work: Ask yourself: is my company/product a factory or a playground? Am I hiring people to execute my instructions, or am I creating conditions where they figure out better paths? Am I producing content for users, or am I creating spaces where users produce value for each other?
4. System Over Individual: The Formation Advantage
Anker: The Big Dipper Formation
Yang's problem: Anker's department heads are "middle horses" — good, not geniuses. How do they beat competitors' CEOs ("top horses") in one-on-one battles?
Answer: don't fight one-on-one. Use a formation.
"In Jin Yong's novels, the Big Dipper Formation uses 7 average martial artists in a coordinated pattern to defeat a master. Wang Chongyang could have fought everyone himself, but he didn't want to — so he built the formation."
Anker's formation has four layers of compounding advantages:
- Channel reuse: Amazon → offline → global, one product supports another
- User group cross-pollination: Security customers buy cleaning, buy storage
- Shared technology: Chips, systems, models — built once, used everywhere
- Unified brand: Anker → eufy → soundcore, one name = one trust signal
Result: the department head isn't alone. They bring the entire formation to every fight.
REMIX: Collaborative IP via Open Source Mechanics
The community framework maps this exactly:
| Open source mechanic | Mapped to collaborative IP |
|---|---|
| Clear main branch | A maintained "canon" version of the IP |
| Forkable + PR-mergeable | Users fork AU (alternate universe) versions; good ones merge back |
| Standardized protocol | Fork/adopt/co-create/consensus mechanism |
| Maintainer role | IP stewards |
| Reputation system | Contributors earn naming rights, revenue share, voice |
The theoretical foundation is Hayek's The Use of Knowledge in Society: modern society works because distributed local knowledge aggregates through exchange. The same applies to creation — IP can be collaboratively grown by distributed contributors.
Fan fiction communities have already proven this works (roles, events, deadlines, recognizable tiers, co-creation mechanisms). AI makes the model massively scalable — characters and worlds no longer degrade as they get remixed.
The shared principle: Don't depend on individual genius. Build a system where coordinated average talent outperforms scattered exceptional talent. The system itself is the moat.
Put it to work: Before hiring a "rockstar," ask: what's the formation they'll fight in? What does your compounding advantage stack look like? If your new person needs to win alone, you have a hiring problem — but also a system problem.
5. Bet Half, Reinvest, Repeat: The Sequencing Principle
Anker: Aggressive Conservatism
"We're aggressive conservatives. Conservative: out of ¥10, we bet at most ¥5-6. Aggressive: we always place that ¥5-6."
The math:
- Start with 10 → Bet 5 → Make 20 → Total 25 → Bet 12 → Repeat
- The pot grows, but the proportion stays at ~50%
This isn't risk aversion. It's sequence discipline: don't over-expand before the system can support it. After 2022's over-expansion (four+ new categories in one year), Yang cut to 10 and rebuilt.
"From 100 products to 30 to ~10... everyone thinks I'm constantly restructuring. Yes."
The sequencing principle also governs who gets paid first:
"In consumer electronics, excess value is created by workers. Workers should get the majority."
| Metric | Before | Now |
|---|---|---|
| Employee comp as % of revenue | ~8% | 13% |
| Employees making >¥1M | 500 (2024) | 800+ (2025) |
| Gross margin | Baseline | +8pp over 5 years |
Compensation structure: distribute opportunity first (values-based), distribute returns after (results-based).
REMIX: Sequence in Community
The playground model has a natural sequencing logic:
Catch a crowd (people who want to remix)
→ Form a community (roles + cultural circle)
→ Build mechanics (fork / adopt / co-create)
→ Let collaboration become assets (IP + collective memory)
→ Product grows into a continuously operating playground
Each step depends on the one before. You can't build mechanics without a community. You can't have a community without a crowd that wants to remix. Skip a step and the system collapses.
The shared principle: Growth isn't about how fast you can go. It's about what you can afford to sustain. Sequence your expansion so each step is funded by the last. Distribute opportunity before returns.
Put it to work: What's your current "pot" (resources, talent, trust)? Never bet more than half. On the next win, reinvest half of the gain. Repeat. And ask yourself: am I distributing opportunity to the right people first, or am I trying to buy my way out of a system problem?
The Unified Chart
| Principle | Anker | REMIX | Get to "yes" question |
|---|---|---|---|
| Terrain over ambition | Shallow sea (≤$50B) vs. deep sea | Cultural boundary vs. mass platform | "Does the shape of my battlefield allow my kind of win?" |
| Ladder over flat | 1357 product hierarchy | Role + hierarchy + sequence | "Do users have a visible path to grow within my system?" |
| Playground over factory | Creator's paradise organization | Open-seed-unit product model | "Am I building a system that enables, or an assembly line that standardizes?" |
| Formation over individual | 4-layer compounding advantage | Forkable IP + maintainer + reputation | "Can 7 average people in formation beat 1 genius alone?" |
| Sequence over speed | Aggressive conservatism (bet 50%) | Step-by-step community building | "Is my next expansion funded by the last win?" |
What This Means for a Founder With 3-24 Months of Runway
This isn't theory. Here's the real checklist:
Month 1: Map your terrain
- Is your category shallow sea or deep sea?
- If deep sea: can you enter through a niche that the dominant player ignores for 3+ years?
- Does your potential community have enough cultural identity to survive growth?
Month 2: Design the ladder
- What is your 1系 (entry, easy to try)?
- What is your 5系 (core, reliable value)?
- What is your 7系 (aspirational, sets the standard)?
- Same question for community: what's the entry role vs. the top role?
Month 3-6: Build the playground, not the factory
- Are you producing things for users, or are you building a space where users produce things for each other?
- What's the "seed unit" — is it completed (factory) or editable (playground)?
- For Anker: the seed unit is a product category operated by a semi-autonomous team. For REMIX: the seed unit is a character/world that can be forked. What's yours?
Month 6-12: Install your formation
- What 3-4 compounding layers can you build that each competitor would need to build from scratch?
- Can your "middle horse" department heads access the same resources that your CEO would?
- For community: can a new contributor become a maintainer through a clear path?
Month 12-24: Sequence, don't sprint
- What is your ¥10? Bet 5. Don't bet 8.
- Did your last win pay for your next expansion?
- Are your most valuable people getting more opportunity, not just more money?
Closing
Yang's ultimate aspiration, quoting Laozi: "The highest ruler is one whose existence is barely known." He places himself at the third level (feared) but wants to reach the first (barely known — the system runs itself).
The community framework says the same thing differently: the best products aren't destinations you visit — they're playgrounds you live in. The creator becomes invisible; the creation becomes self-sustaining.
A company is the hardest product. A community is the hardest company. But the principles that govern them — terrain, ladder, playground, formation, sequence — are the same.
The question isn't what you build. It's what system you set in motion.
Top comments (0)