What Kinds of Businesses Are There?
Re-understanding the underlying structure of different businesses through four dimensions: product evolution, customer accumulation, core assets, and revenue compounding.
Here is the conclusion up front:
This classification method is, at its core, not a division by “industry,” but a division by whether cash flow accumulates, whether the product keeps evolving, whether customers compound, and whether assets can be replicated.
The true underlying structure of any venture can be judged with these four questions:
1. Will the product keep changing?
Is it the “finite-improvement type” or the “infinite-improvement type”?
2. Will customers accumulate?
Is it the “well-digging type” or the “reset-to-zero type”?
3. Where do the core assets accumulate?
In brand, channels, physical locations, technology, data, licenses, IP, organizational capability — or in personal experience?
4. Can revenue compound?
Is it one-off transactions, repeat purchases, subscriptions, commissions, rent, service fees, royalties, or capital gains?
-–
I. The First Layer of Classification: Finite-Improvement vs. Infinite-Improvement
1. Finite-improvement ventures
The defining feature of this kind of business:
Once the product matures, its core form no longer changes fundamentally; subsequent improvements are mainly about packaging, channels, cost, efficiency, and brand expression.
Typical examples:
Type Examples Underlying logic
Beverages Coca-Cola, Nongfu Spring, C’estbon Taste and demand are stable; the core is brand, channels, coolers, and store-level coverage
Condiments Haday soy sauce, Lee Kum Kee, Lao Gan Ma Consumer habits are stable and product improvement is limited; the core is trust and repeat purchase
Household and personal care Toothpaste, laundry detergent, tissue Functions are mature; success rests mainly on brand mindshare, channel efficiency, and supply-chain cost
Tobacco and liquor Chunghwa cigarettes, Moutai, Wuliangye The core product is stable; consumption occasions and status value matter more
Staple foods Rice, flour, oil, salt, sugar The product changes little; the game is won on scale, channels, cost, and consistent quality
Its core is not “innovation,” but:
> Selling a mature product, steadily and over the long term, to a large number of people.
What matters most in this kind of business:
Brand trust;
Channel density;
Supply-chain efficiency;
Cost control;
Share of retail outlets;
Stability of consumption habits.
Once such a business is established, cash flow is strong, but the downsides are that growth tends to slow, younger consumers may drift away, and channel shifts can erode the original advantages.
-–
2. Infinite-improvement ventures
The defining feature of this kind of business:
The product itself keeps evolving; user needs, technological conditions, and competitive standards are all in flux. You cannot stop — the moment you stop, you fall behind.
Typical examples:
Type Examples Underlying logic
Smartphones iPhone, Huawei, Xiaomi Hardware, operating systems, ecosystems, AI, and imaging iterate continuously
New-energy vehicles Tesla, BYD, Li Auto Batteries, intelligent driving, cabins, and charging systems keep evolving
Software / SaaS Feishu (Lark), Notion, Salesforce Features, collaboration, AI capabilities, and ecosystem plugins are continuously updated
AI tools ChatGPT, Claude, Workbuddy Models, agents, workflows, and data integrations keep upgrading
Convenience stores 7-Eleven, Lawson, regional chains Product mix, fresh food, membership, instant retail, and supply chains are continuously optimized
E-commerce platforms Taobao, JD.com, Pinduoduo, Meituan Traffic rules, fulfillment, algorithms, subsidies, and merchant systems change constantly
Its core is not “defending one thing,” but:
> Continuously learning, continuously iterating, continuously adapting to new needs.
The advantage of this kind of business is ample room for growth; the downside is that you can never coast — R&D, organization, data, and product capability all demand continuous investment.
-–
II. The Second Layer of Classification: Well-Digging vs. Reset-to-Zero
1. Well-digging businesses
The defining feature of a well-digging business:
Acquire a customer once, and that customer keeps contributing revenue.
Take the third-party labor outsourcing you mentioned: once a company hands its social-insurance filings, payroll, and workforce settlement over to you, as long as nothing goes wrong, you can keep charging month after month.
Typical examples:
Type Examples Revenue logic
Labor outsourcing Third-party staffing, flexible staffing, social-insurance agency services Sign once, collect service fees long term
Property management Residential compounds, business parks Enter once, keep collecting management fees
SaaS Enterprise software, ERP, CRM Subscription fees, module fees, value-added services
Insurance Auto, life, commercial insurance Renewed year after year
Bank custody Corporate accounts, funds settlement Deposit float, transaction fees, interest spread
Payment acquiring WeChat Pay, Alipay, POS service providers A commission on every transaction
Convenience-store locations Community, school, and hospital stores Once the location is fixed, nearby foot traffic keeps flowing in
The key variable in a well-digging business:
> Customer-acquisition cost can be amortized over the long term.
What it values most:
Retention rate;
Renewal rate;
Customer switching costs;
Service stability;
Depth of lock-in;
Customer lifetime value.
This kind of business is not necessarily wildly profitable, but it is well suited to long-term operation.
-–
2. Reset-to-zero businesses
The defining feature of a reset-to-zero business:
Finish this one, and next time you start all over again.
Typical examples:
Type Examples Why it resets to zero
Training Corporate training, open-enrollment courses, industry courses Once this session ends, the next has to be sold all over again
Consulting Strategy consulting, project consulting When the project ends, the revenue ends
Film Every film is greenlit, shot, and marketed from scratch The last film’s success does not guarantee the next one’s
Advertising Annual campaigns, brand planning Every project is re-bid
Conferences and exhibitions Industry summits, launch events, expos Every edition means re-selling sponsorships and re-organizing
Renovation Home renovation, commercial space design One-off transactions, low-frequency repurchase
Custom development Software outsourcing, system development Project-based delivery — once finished, it’s over
Reset-to-zero businesses are not bad — they are just stressful.
Their problem:
> Revenue does not roll forward on its own; you must keep selling, keep creating, keep delivering.
So the most important move in a reset-to-zero business is to convert one-off projects into assets that accumulate, for example:
Turning training into long-term memberships;
Turning consulting into annual advisory retainers;
Turning courses into certification systems;
Turning films into franchise IP;
Turning project experience into methodology;
Turning custom development into standardized software;
Turning client relationships into long-term service contracts.
-–
III. Combining the Two Dimensions Yields Four Basic Types of Venture
Combination Representative type Underlying logic Examples
Finite improvement + well-digging The steadiest cash-flow businesses Stable product, customers keep repurchasing Coca-Cola, Nongfu Spring, property management, basic condiments
Finite improvement + reset-to-zero Low-frequency mature categories Mature product, but customers don’t persist Furniture, renovation, weddings, some home appliances
Infinite improvement + well-digging High-growth compounding businesses Product keeps upgrading, customers stay bound in The iPhone ecosystem, SaaS, cloud services, convenience-store chains, labor-outsourcing platforms
Infinite improvement + reset-to-zero High-volatility creative businesses Must innovate every time, must re-sell every time Film, training, consulting, advertising, fashion, the content industry
The most comfortable is the first type: finite improvement + well-digging.
The highest-growth is the third type: infinite improvement + well-digging.
The most exhausting is the fourth type: infinite improvement + reset-to-zero.
But if a fourth-type business builds IP, a brand, memberships, or systems, it can turn into the third or the first type.
-–
IV. The World’s Main Business Archetypes
What follows is grouped not by industry, but by underlying structure.
-–
1. Formula businesses
Representatives: cola, soy sauce, baijiu, toothpaste, tissue, bottled water.
Underlying logic:
> A stable product consumed at high frequency, with long-term repurchase driven by brand and channels.
Core assets:
The formula;
Brand;
Channels;
In-store display;
Consumer habits;
Supply-chain scale.
The key to this kind of business is not “inventing new products every year,” but getting consumers to form habits.
What Coca-Cola really sells is not a bottle of sugar water, but a global brand, channels, coolers, consumption occasions, and emotional memory.
-–
2. Iterative product businesses
Representatives: smartphones, cars, software, AI tools, smart hardware.
Underlying logic:
> User needs keep escalating, old products keep depreciating, and the company must keep iterating.
Core assets:
R&D capability;
Product capability;
User-feedback systems;
Supply-chain coordination;
Operating systems;
Ecosystem capability.
The smartphone is the classic case.
The first generation of smartphones solved “it works”; later generations solved “it works well”; later still came photography, payments, entertainment, office work, and AI assistants.
With every round, users’ standards rise — which is what makes this the infinite-improvement type.
-–
3. Subscription businesses
Representatives: SaaS, membership programs, cloud services, video-streaming memberships, paid-knowledge memberships.
Underlying logic:
> Users don’t buy the product once; they keep buying the right to use it.
Core assets:
Retention;
Renewals;
Account systems;
Accumulated data;
Feature iteration;
Customer-success systems.
The advantage of subscription businesses is predictable revenue.
But they carry one pressure: every month and every year, users are re-deciding whether renewal is “worth it.”
-–
4. Taxation / commission businesses
Representatives: payments, platform transactions, food-delivery platforms, online travel agencies, stock exchanges, the App Store.
Underlying logic:
> They don’t produce goods directly; they control the transaction channel and take a small cut of every transaction.
Core assets:
Transaction gateways;
User scale;
Merchant scale;
Payment infrastructure;
Trust mechanisms;
Rule-setting power.
Once established, this kind of business is extremely strong.
Because the more others transact, the more you earn.
But its difficulty lies in the early cold start and in platform governance.
Without users, merchants won’t come; without merchants, users won’t come either.
-–
5. Rent-collecting businesses
Representatives: commercial real estate, storefronts, parking lots, billboards, warehouses, server rooms.
Underlying logic:
> Control scarce space or scarce assets, and rent out the right to use them.
Core assets:
Location;
Property rights;
Leases;
Foot traffic;
Stable occupancy;
Asset appreciation.
This business is, in essence, “asset possession.”
Its advantage is stability; its drawbacks are heavy assets, long cycles, and high exposure to urban change and interest rates.
-–
6. License businesses
Representatives: banking, insurance, tobacco, healthcare, education credentials, city gas, highways, ports, mines.
Underlying logic:
> Not everyone can get in; the right of entry is itself the moat.
Core assets:
Policy permits;
Licenses;
Franchise and concession rights;
Control of resources;
Compliance capability.
Profits in this kind of business come from scarce access.
But its risk also comes from policy: once the regulatory logic changes, the business model gets rebuilt.
-–
7. Platform businesses
Representatives: Taobao, Meituan, Didi, Douyin, WeChat, Airbnb.
Underlying logic:
> Connect multiple sides, lower matching costs, and build network effects.
Core assets:
The user network;
The merchant network;
Algorithmic matching;
Data;
Payments and fulfillment;
Platform rules.
The essence of a platform business is not selling goods, but building a marketplace.
Once the marketplace takes shape, the platform holds pricing power, distribution power, and rule-making power.
-–
8. IP businesses
Representatives: Disney, Marvel, Harry Potter, Pokémon, Ultraman, game characters, web-novel IP.
Underlying logic:
> The initial creation is reset-to-zero, but once the IP succeeds, it can be monetized again and again.
Monetization channels:
Films;
Games;
Licensing;
Toys;
Theme parks;
Co-branded merchandise;
Membership content.
IP businesses have a dual structure:
> The creation phase is reset-to-zero; the monetization phase is well-digging.
That’s why a standalone film is dangerous, but the Marvel universe is different.
It doesn’t start from zero each time; it repeatedly harvests attention and emotional assets within the same fictional world.
-–
9. Expert businesses
Representatives: lawyers, doctors, consultants, designers, investment-banking advisors, trainers.
Underlying logic:
> What customers buy is a person’s judgment, experience, taste, credibility, and problem-solving ability.
Core assets:
Professional reputation;
Case portfolio;
Methodology;
Networks;
Trust;
Judgment.
The problem with expert businesses is that they are easily capped by individual capacity.
One expert can serve only so many clients in a day.
So for an expert venture to scale, it must complete three upgrades:
1. From personal experience to methodology;
2. From methodology to team processes;
3. From team processes to products, systems, or tools.
-–
10. Content / attention businesses
Representatives: independent creators, multi-channel networks, livestreaming, short-form dramas, news media, variety shows, stand-up comedy.
Underlying logic:
> Gather attention, then monetize it through advertising, e-commerce, memberships, tipping, and IP licensing.
Core assets:
Traffic;
Persona;
Content-production capability;
The account matrix;
Fan relationships;
Understanding of distribution platforms.
This kind of business resets to zero very easily.
One viral post doesn’t guarantee the next; one blockbuster livestream doesn’t mean the next will hit too.
But if it develops into ongoing programs, IP, communities, memberships, courses, and brand partnerships, it migrates from the reset-to-zero type toward the well-digging type.
-–
11. Supply-chain-efficiency businesses
Representatives: Costco, JD Logistics, SHEIN, regional B2B, convenience-store supply chains, integrated warehousing-and-delivery.
Underlying logic:
> Not creating new demand, but doing procurement, inventory, turnover, fulfillment, and cost better than everyone else.
Core assets:
Procurement scale;
The warehousing-and-delivery network;
Inventory management;
Turnover efficiency;
Systems capability;
Upstream and downstream relationships.
Margins in this kind of business are usually thin, but once scale is reached, it is very stable.
Its advantage is not high gross margin, but high efficiency, fast turnover, low shrinkage, and good cash flow.
-–
12. Data / algorithm businesses
Representatives: recommendation systems, risk-control systems, site-selection models, BI systems, ad-targeting algorithms, the DSCM product-selection system.
Underlying logic:
> More data means sharper judgment; sharper judgment means higher business efficiency.
Core assets:
Data sources;
Tagging taxonomies;
Model capability;
Closed feedback loops;
Understanding of the business scenario;
Decision systems.
The key to this kind of business is not “having AI,” but:
> Whether you have real business data, whether you can close the loop, and whether you can keep calibrating.
A data system without a closed loop easily becomes a system for show;
only a data system with a closed loop becomes a system that runs the business.
-–
13. Finance / capital businesses
Representatives: banks, insurance, funds, VC, PE, financial leasing, factoring, microlending.
Underlying logic:
> Making money from capital, risk pricing, and time differentials.
Core assets:
Cost of capital;
Risk identification;
Credit systems;
Leverage control;
Investment judgment;
Exit capability.
Banks earn the interest spread;
insurers earn from the risk pool and the time gap;
VCs earn from a few big wins covering a large number of failures.
Finance businesses look like “money making money,” but their essence is:
> Managing risk with judgment, and trading time for returns.
-–
14. Community / membership businesses
Representatives: Costco, Sam’s Club, the knowledge platform Dedao, private-domain communities, clubs, religion-like brands, industry associations.
Underlying logic:
> First build identity and trust, then keep providing products, content, services, and transaction opportunities.
Core assets:
Member relationships;
Sense of identity;
Trust;
High-quality supply;
Organized activities;
A sustained sense of value.
The key to this kind of business is not herding people into group chats, but making users feel:
> I belong here, and this place keeps being useful to me.
-–
15. Standard / protocol businesses
Representatives: iOS, Android, Visa, the USB standard, operating systems, enterprise-software ecosystems, AI agent protocols.
Underlying logic:
> Not just selling a product, but defining how others connect, develop, transact, and collaborate.
Core assets:
Technical standards;
The developer ecosystem;
Interfaces;
Protocols;
Rules;
Control of the ecosystem’s key positions.
Once established, this kind of business is extraordinarily strong.
Because others aren’t simply buying your product — they are living inside your rules.
-–
V. A Few Typical Cases, Taken Apart
1. Coca-Cola
It is not simply a beverage business, but:
A finite-improvement product;
High-frequency repurchase;
A global brand;
Channel control;
Occupation of consumption occasions;
Emotional assets.
So the core of Coke is not “constant product innovation,” but:
> Getting a stable product consumed over and over, across countless occasions around the world.
-–
2. Apple
Apple is not a pure smartphone company, but a composite venture:
Phone hardware: infinite-improvement type;
The iOS system: standard/protocol type;
The App Store: commission type;
iCloud and Apple Music: subscription type;
The brand: premium-mindshare type;
The ecosystem: well-digging type.
Apple’s strength is that it doesn’t just sell you a phone — it places users inside a long-term ecosystem.
-–
3. Disney
Disney is the classic combination of “reset-to-zero creation + well-digging IP.”
Animation and films: every creation carries reset-to-zero risk;
Mickey Mouse, Marvel, Frozen: reusable IP assets;
Theme parks: rent-collecting / experience type;
Licensed merchandise: royalty type;
Streaming: subscription type.
Its real genius lies in turning a single act of creation into a long-term asset.
-–
4. Convenience stores
The convenience store is extremely complex — it is not a single business.
It simultaneously combines:
Layer Attribute
Location Well-digging type; nearby foot traffic keeps flowing in
Merchandise Finite improvement + high-frequency repurchase
Assortment Infinite improvement; must continuously adapt to the trade area
Fresh food Infinite improvement; demanding on shrinkage and supply chain
Membership Data type, repurchase type
Instant retail Platform-competition type
Supply chain Efficiency type
Franchise system Standardized-replication type
So the difficulty of the convenience store lies here:
> It looks like opening stores, but it is actually a composite system of locations, merchandise, supply chain, membership, data, and organizational processes.
A single-store owner can survive on experience;
a convenience-store chain survives on systems, supply chain, and operating models.
-–
5. A business like New Distribution
New Distribution (Xinjingxiao) naturally carries several attributes today:
Attribute How it shows up
Content type Articles, reports, industry viewpoints
Expert type Judgment, consulting, speaking, training
Reset-to-zero type Every training session, project, and event must be sold anew
IP type The New Distribution brand and its industry influence
Community type Relationships among distributors, brands, and retailers
Upgrade paths Memberships, course systems, data products, AI workbench, skill packs, advisory subscriptions
So if New Distribution only does training, consulting, and events, it skews toward the reset-to-zero type.
If it upgrades in the following directions, it becomes a well-digging business:
Annual memberships;
An industry database;
An AI workbench for distributors;
A standardized course system;
Corporate advisory subscriptions;
Skill packs for convenience stores and distributors;
Systems for product selection, ordering, and business diagnosis;
Industry certification and training systems.
In other words, New Distribution’s real strategic direction is not “run a few more training sessions,” but:
> Turning one-off knowledge services into ongoing operating infrastructure.
-–
VI. To Judge Whether a Business Is Good, Look at This Table
Dimension Traits of a good business Traits of a poor business
Customer relationships Customers keep repurchasing Every sale requires new acquisition
Product form Standardized, replicable Highly customized, delivery-heavy
Revenue structure Subscriptions, repurchases, commissions, rent One-off project income
Cost structure Marginal costs fall Labor costs rise in lockstep with revenue
Moat Brand, channels, data, licenses, networks Only the owner’s personal ability
Cash flow Stable, predictable Up one moment, down the next
Improvement pressure Controlled iteration Perpetual firefighting
Scalability Replicable Grows only by piling on headcount
Asset accumulation Every delivery leaves an asset behind Once finished, nothing remains
Risk sources Manageable Dependent on hits, connections, and luck
-–
VII. The One Sentence Most Worth Remembering
Every venture can be asked one question:
> Will what you do today become an asset tomorrow?
If yes, it is a compounding venture.
If not, it is a consuming venture.
Reset-to-zero businesses are not off-limits, but you must find a way to convert them into accumulated assets.
Finite-improvement businesses are not lowly; they earn their compounding through brand, channels, and scale.
Infinite-improvement businesses are not superior; the organization must keep evolving, or it will quickly be eliminated.
Well-digging businesses are the most valuable — provided the well actually holds water, and customers won’t easily switch wells.