You know you want to start something. What you don’t know is what kind of thing it should be. A product? A service? A business built on what you know? The answer changes almost everything that follows: how much money you need, how fast you can make a first sale, and how hard it is to prove anyone wants it.
This guide sorts the options into four types and helps you pick the one that fits what you have to offer. If what you’re selling is your knowledge, one path is far lighter to start than the others.
The four ways to start a business
Nearly every small business is one of four types. They differ less in industry than in what the customer is paying for, and that difference drives the cost, speed, and risk.
The Four Ways to Start a Business
Compare what customers buy, what it takes to begin, and how each model grows.
| Dimension | Physical product | Traditional service | Expertise-led | Knowledge product |
|---|---|---|---|---|
| What customers buy | A physical object | A task done for them | Your judgment on their problem | A packaged thing they use without you |
| Time to first sale | Slow: source, make, stock | Fast: book the job | Fast: a conversation | Slow: build it, find an audience |
| Upfront cost | High: inventory, materials | Low to moderate: tools, licensing | Very low: your time | Low to moderate: time to build |
| Distribution | Retail, marketplaces, shipping | Local reach and referrals | Relationships and reputation | An audience you can market to |
| Delivery complexity | High: make, ship, returns | Moderate: your hands, on site | Moderate: per client, your input | Low per sale once built |
| Evidence needed | People buy above your cost | Local demand for the task | Someone pays for your judgment | People pay for the package |
| Scalability | High, but capital-heavy | Limited by your hours | Moderate; grows once packaged | High once it sells |
| LookSky fit | Not a fit | Partial | Strong | Strong |
Physical product
- Customer buys
- A physical object
- First sale
- Slow: source, make, stock
- Upfront cost
- High: inventory and materials
- Distribution
- Retail, marketplaces, shipping
- Delivery
- High complexity
- Evidence
- People buy above your cost
- Scale
- High, but capital-heavy
- LookSky fit
- Not a fit
Traditional service
- Customer buys
- A task done for them
- First sale
- Fast: book the job
- Upfront cost
- Low to moderate
- Distribution
- Local reach and referrals
- Delivery
- Your hands, often on site
- Evidence
- Local demand for the task
- Scale
- Limited by your hours
- LookSky fit
- Partial
Expertise-led
- Customer buys
- Your judgment on their problem
- First sale
- Fast: a conversation
- Upfront cost
- Very low: your time
- Distribution
- Relationships and reputation
- Delivery
- Per client, with your input
- Evidence
- Someone pays for your judgment
- Scale
- Grows once you package it
- LookSky fit
- Strong
Knowledge product
- Customer buys
- A packaged thing used without you
- First sale
- Slow: build it, find an audience
- Upfront cost
- Low to moderate time cost
- Distribution
- An audience you can market to
- Delivery
- Low per sale once built
- Evidence
- People pay for the package
- Scale
- High once it sells
- LookSky fit
- Strong, usually later
Read down the expertise-led column: very low cost, fast to a first sale, and the cleanest evidence—a real person paying for your judgment. That’s why it’s a strong starting point for many experienced people.
Physical-product businesses
If you sell an object, it has to exist before anyone can buy it: sourcing or making it, holding inventory, handling shipping and returns. These can be excellent businesses, but they ask for money and logistics up front, before you know whether the thing sells.
Product businesses tend to require real money up front, because inventory and equipment cost real money. It’s the heaviest of the four paths, and not what LookSky is for.
Traditional service businesses
Here you perform a task: cleaning, landscaping, repair, catering, personal training. Customers pay for the work, usually in person and often locally. You can book a first job fast with modest tools. The limit is your own hours; to grow past what you can personally do, you hire and manage people.
An expert who advises a restaurant on its operations is running an expertise-led business. The person who cleans that restaurant is running a traditional service business. Both are real and valuable; they just start and scale differently.
Expertise-led businesses
This is often the lightest path, and the one this whole site is built around. You take the judgment you’ve built over years and apply it to a customer’s problem. Consulting, advisory work, fractional roles, coaching, and done-for-you expert services all live here. Customers aren’t really buying your hours; they’re buying the result your experience produces.
Why it’s often one of the lightest to start
- Almost no upfront cost. Your main input is what’s already in your head.
- A fast first sale. A paying client can come from a few conversations, not a warehouse.
- The cleanest possible evidence. When someone pays you to solve their problem, you’ve proven demand directly.
If you’re still working out what your sellable expertise even is, start with what you can sell from what you already know.
Knowledge products
A knowledge product is what you know, packaged so people can buy it without you: a course, a guide, a template kit, a paid community. Once built, each sale takes little time and can scale to many buyers. That’s the appeal.
The catch is the order. A knowledge product is the hardest of the four to prove early, because you build it first and then have to find enough of the right buyers. Plenty of good courses die not because the knowledge was bad, but because they were built before anyone confirmed people would pay for that exact package.
The one distinction that saves you months
The most common mistake experienced people make is jumping straight to the knowledge product—deciding the business is a course before they’ve sold their expertise to a single person. Do it the other way around.
Sell your judgment directly first, one client at a time. That teaches you something a course can’t: the exact problem people keep coming back for, in the words they use, at a price they’ll pay. After you’ve solved it for enough people, the pattern is obvious, and now you know precisely what to package.
Sell direct first. Package later.
The first customers are not a detour. They are the research.
Sell direct expertise
Help real clients with real problems, hands-on.
Watch for the repeat
Notice the problem you keep getting hired to solve.
Package it later
Turn the proven, repeated solution into a product.
Selling direct first is the research that makes a product work.
Selling direct first isn’t a lesser version of the course. It’s the research that makes the course work, and the cheap way to confirm demand before you build—which is the whole point of knowing if people will pay before you build a website.
Which one is for you?
Registering a company sits alongside all of this, not before it. The SBA notes that your business structure affects taxes, liability, and paperwork, and that rules vary by state. Since rules differ by location, treat this as general information and check an official source like the SBA or IRS for your situation. And remember: a registration, a logo, and a bank account are not proof anyone will buy. A paying customer is.
SAM’S FIELD NOTE
I’ve started on both sides of this decision
“If what you’re selling is what you know, you can start far lighter than I did with clothes.”
I’ve started three companies. Two of them, 9Tailors and Grae Cove, sold physical things: custom clothing and linen. I loved that work, but I won’t pretend it was light. Before I could sell anything, there was fabric to source, inventory to hold, samples to make, orders to fulfill, and returns to handle.
LookSky came from noticing a different starting point. Plenty of experienced people are sitting on something they can sell without any of that: their judgment, the thing colleagues already come to them for. That business can begin with a buyer, a problem, and a few conversations. No warehouse required.
Physical-product and service businesses aren’t worse; they’re just different, each with its own pros and cons. But if what you’re selling is what you know, you can start far lighter than I did with clothes. That lighter path is the one this whole site is about.
Frequently asked questions
What kind of business should I start if I have expertise but no product?
Start an expertise-led business, where customers pay for your judgment on their problem rather than for an object. It’s often one of the lightest of the four types to launch: low upfront cost, a first sale from a few conversations, and clean proof of demand the moment someone pays you. Consulting, coaching, and fractional roles all fit here.
What’s the difference between a service business and an expertise-led business?
A traditional service business sells a hands-on task, often done locally and in person, like cleaning or repair. An expertise-led business sells your judgment applied to a client’s problem, like advising that same business on its operations. The task business is limited mostly by your hours; the expertise business grows when you package what you know.
What is an expertise-led business?
An expertise-led business is one where customers pay for your knowledge, judgment, and the results they produce, not for a physical product or a manual task. Examples include consulting, advisory work, fractional roles, coaching, and done-for-you expert services. It’s often one of the lightest types for an experienced professional to start, because your main asset is what you already know.
Should I create an online course to sell what I know?
Not as your first move. A course is a knowledge product, the hardest type to prove early because you build it before you know it sells. Sell your expertise directly to clients first, notice the exact problem people keep hiring you for, then package that proven solution into a course once demand is clear. The direct work is your research.
Do I need a physical product or inventory to start a business?
No. A business can exist without any employees at all: the Census Bureau counted 30.4 million American businesses with no employees in 2023. If what you offer is your knowledge and judgment, you can run a real business with no inventory, warehouse, or manufacturing, selling your expertise rather than a physical product.
Is consulting a good first business for an experienced professional?
Yes, for many people it’s the best first business. Consulting is an expertise-led business, so it starts cheaply, reaches a first paying client quickly, and proves demand the moment someone pays for your judgment. It also teaches you the repeated problem buyers care about, which is what you’d need before packaging anything into a course later.
How much money do I need to start an expertise-led business?
An expertise-led business can often begin with low upfront costs, because it needs no inventory or manufacturing and your main input is knowledge you already have. Actual costs still vary by location and business: registration, insurance, software, and professional advice all differ. Early spending is usually modest, but check your own local requirements rather than assuming a fixed figure.
Do I need to register a company before I start selling?
Rules vary by location, so check an official source, but registering a company is a separate legal step from proving demand. The SBA notes your business structure affects taxes, liability, and paperwork and should be chosen before you register with your state. A registration, logo, or bank account doesn’t show anyone will buy. Validate demand through real conversations first.
Sources
- U.S. Census Bureau, Nonemployer Statistics— supports the 30.4 million businesses with no employees figure for 2023.
- U.S. Small Business Administration, Choose a business structure— general guidance on taxes, liability, paperwork, and registration.
The comparison matrix and “sell direct first, package later” sequence are LookSky editorial frameworks, not measured performance data.