How Much Does It Actually Cost to Start a Business?

Every answer to this question is either a fantasy or a shrug. The fantasy is "a hundred dollars and a laptop." The shrug is "it depends." Neither one helps you decide whether to quit your job.

So here are the numbers. Real ranges, broken out by business type, with the variables that explain the gap between the low end and the high end — and the costs that never appear on anyone's startup checklist until they appear on your bank statement.

Why "it depends" is a useless answer

It does depend. That is not the problem. The problem is that "it depends" is where most advice stops, when it should be where the advice starts.

Almost all of the variance in startup cost comes down to three questions:

1. Does the business need inventory or equipment before the first sale? Physical products and trades do. Services do not.
2. Does the thing you sell have to be built before anyone can buy it? Software does. Consulting does not.
3. How much of the work are you doing yourself? The same business costs $3,000 or $25,000 depending on whether you build the website and write the copy or pay someone to.

A business that answers no, no, and "myself" starts for under $2,000. A business that answers yes, yes, and "someone else" starts north of $75,000. Every business you can name sits somewhere on that grid, and once you know where yours sits, the range narrows fast.

The other reason people give vague answers is that they are quietly mixing two different things: the cost to open the doors, and the cost to stay alive until the business pays you. The second number is usually larger, and it is the one that kills businesses. We will get to it.

The headline numbers

This is setup cost only — everything you spend to be legally able to take money from your first customer. It does not include your salary, your rent, or the months of losses before the business turns over. Those are further down.

Two things to notice.

The lean column is not fictional, but it assumes you do nearly everything yourself and buy nothing you can borrow. The SaaS lean number of $2,000 assumes a founder who writes the code. If you cannot write the code, that column does not exist for you, and pretending otherwise is how people end up $40,000 into a product nobody asked for.

The spread inside a single row is often 10x, and it is almost never about quality. It is about impatience. The expensive version of a local service business is the one that buys a wrapped van in month one instead of month nine.

Where the money actually goes

Local service business

The trade determines the equipment, and the equipment determines the number.

• Entity and licensing: $150 – $1,200. State LLC filing is $50 – $500 depending on the state, most commonly $100 – $200. A local business licence adds $50 – $400. Regulated trades add a contractor licence, exam fees and often a surety bond — the bond premium is typically $100 – $500 a year on a $10,000 – $25,000 bond.
• Insurance: $600 – $2,500 a year for general liability on a low-risk trade. Higher-risk work such as roofing or tree removal runs several times that. Commercial auto is $1,200 – $3,000 per vehicle per year and is not optional the moment you use a personal truck for work.
• Vehicle: $0 if you already own something suitable. A used work van or truck is $8,000 – $25,000; new is $35,000 – $60,000; leasing is $400 – $900 a month and converts a capital cost into a fixed monthly one, which matters more than the total.
• Equipment: $500 – $2,000 for cleaning, $3,000 – $8,000 for a pressure-washing rig, $5,000 – $15,000 for a lawn setup, more for anything mechanical.
• Getting found: a Google Business Profile is free and is the highest-return hour you will spend. A website is $0 – $500 if you build it or $2,000 – $8,000 from a freelancer. Vehicle lettering is $300 – $800; a full wrap is $2,000 – $4,000 and can wait.

Total for a genuinely lean start where you already own the vehicle: $2,000 – $5,000. Add a vehicle and you are at $15,000 – $30,000 quickly.

Ecommerce brand

This is where budgets go wrong, because the inventory number is only the first inventory number.

• Product development and samples: $200 – $1,500 across two or three sample rounds. Custom tooling or moulds, if your product needs them, is $1,000 – $10,000 on its own.
• First inventory order: $2,000 – $25,000. Minimum order quantities do most of the work here — a factory asking for 1,000 units at $5 landed is a $5,000 cheque before you have sold one.
• Packaging: design and dielines $500 – $3,000, then printed boxes or mailers at $1 – $4 a unit with their own minimums.
• Photography: $500 – $3,000 for a starter set. Roughly $50 – $150 per product for clean studio shots, more for lifestyle.
• Store: Shopify Basic is $39 a month, the mid tier around $105, and apps quietly add $50 – $300 a month. A designed store on top is $2,000 – $15,000.
• Freight and duties: sea freight is far cheaper per unit than air but adds thirty to sixty days, which is a cash-flow cost rather than a line item. Duty rates depend entirely on your product classification and origin, and can range from zero to well over 20 percent. Model it before you order, not after.
• Ads to find out whether it sells: $2,000 – $5,000 minimum to get a read you can trust. Less than that and you are reading noise.

The trap: the second inventory order comes due before the first has sold through. A brand that spends its last $8,000 on inventory and sells it in ninety days has no cash to reorder and dies profitable on paper.

Productized service

A fixed offer at a fixed price, delivered the same way every time. It is the cheapest way to build something that is not just you selling hours, which is why it is worth understanding before you pick a model — see choosing a business model.

• Entity and insurance: $600 – $2,000 a year including professional liability.
• Offer definition and positioning: the actual work of deciding what you sell and what you refuse to sell. Free if you do it, $1,500 – $6,000 if you want help.
• Landing page and intake: $500 – $5,000. The intake form and the delivery checklist matter more than the design.
• Delivery capacity: the real cost. If you subcontract fulfilment, you need working capital to pay your contractor before the client payment clears — budget one full delivery cycle, commonly $1,000 – $5,000.
• Automation: $50 – $300 a month of tooling to keep intake, scheduling and follow-up from eating the margin.

SaaS

The cheapest infrastructure and the most expensive labour in this list.

• Infrastructure: genuinely $50 – $500 a month early on. Hosting, database and email are not what makes software expensive.
• Building the first version: $0 if you write it, and six to twelve months of your life. Contract development for a real first version with accounts, billing and a working core is $15,000 – $60,000. Anything with multi-tenancy, integrations, permissions or compliance runs $60,000 – $250,000.
• Billing: Stripe Billing adds roughly 0.5 percent on top of standard processing for recurring charges.
• Compliance, when enterprise buyers start asking: SOC 2 Type II typically costs $20,000 – $60,000 in the first year once you add the auditor, the compliance tooling and the engineering time to close gaps. Not a day-one cost, but do not be surprised by it.
• Support and monitoring tooling: $100 – $500 a month.

If you are non-technical, the honest SaaS number is the contract development number, and the single most expensive mistake in this whole article is paying it before anyone has agreed to buy. Validate the idea first — a landing page and twenty sales conversations cost a few hundred dollars and answer the same question.

Professional service

The cheapest business type to start and the one most often over-built.

• Entity, insurance, accounting setup: $1,000 – $3,000 in year one, including errors-and-omissions cover at $500 – $1,500 a year.
• Regulated professions add licensing and continuing education at $200 – $2,000 a year.
• Website and positioning: $0 – $8,000. Everything above the low end buys credibility, which for this business type genuinely converts.
• Everything else is optional for longer than people think. Office space, staff and software all come after the client list, not before it.

The costs nobody puts on the list

These are annual figures for a small business doing real but modest revenue. They are the difference between a budget that survives contact with reality and one that does not.

A few of those deserve more than a row.

Sales tax is now a multi-state problem, not a home-state one. Since the 2018 Wayfair decision, states can require you to collect based on economic activity alone. The most common threshold is $100,000 of sales into a state, with some states also counting transactions. Cross it and you are registering, filing and remitting in that state — often monthly. For a service business this rarely bites. For ecommerce it eventually always does, and the software to handle it starts around $50 a month and climbs with your state count.

The software stack is death by $29 a month. Email at $7 – $22 per user, accounting at $35 – $99, a CRM at $0 – $150 per user, email marketing that scales with your list, scheduling, storage, design tools, hosting. A modest small business commonly runs $200 – $500 a month, and almost nobody budgets for it because each individual subscription feels trivial.

Payment processing is a tax on the whole business

Standard online card processing is 2.9 percent plus 30 cents per transaction. In person it is nearer 2.6 percent plus 10 cents. International cards add roughly 1.5 percent, currency conversion adds about 1 percent, and a disputed charge costs around $15 whether or not you win.

The fixed per-transaction fee is what people miss. Work it out:

• $200,000 of revenue at an average order of $50 is 4,000 transactions. The percentage fee is $5,800, the per-transaction fee is $1,200, so you pay $7,000 — an effective 3.5 percent.
• The same $200,000 at an average order of $500 is 400 transactions. You pay $5,800 plus $120, or 2.96 percent.

Same revenue, $1,080 difference, entirely explained by order size. If your average order is under $25, card fees are a serious line item and worth designing around with bundles, minimums or ACH — bank transfers are commonly under 1 percent with a hard cap, which for B2B invoices is free money.

Your own time is the largest line item and it never appears

If you spend six months building a business while forgoing a $70,000 salary, the business cost $35,000 before a single invoice. That is not an accounting fiction. It is the number you will compare against when you are tired in month eight.

You do not have to pay yourself. You do have to write the number down, for two reasons. It stops you treating your own labour as free and therefore infinite — which is how founders end up doing $20 an hour work on a business that needs $200 an hour decisions. And it puts a ceiling on how long you are willing to stay unprofitable, decided while you are calm rather than while you are desperate.

A workable rule: assign yourself a modest hourly figure, track roughly what you spend, and any time a task costs less to buy than to do, buy it. Bookkeeping is the usual first candidate.

What you can genuinely defer

The pattern: defer anything whose value depends on being at scale, and buy anything that protects you from a single bad day.

What is a false economy

Some cheap decisions are genuinely expensive.

• Skipping insurance. One liability claim, one damaged floor, one client who says your advice cost them money, and the $900 a year you saved becomes the least relevant number in the room.
• Mixing personal and business money. A separate bank account costs nothing. Untangling eighteen months of mixed transactions costs $1,500 – $4,000 in bookkeeper time and puts your liability protection in question.
• No bookkeeping until tax season. Reconstruction is billed at a premium, you lose deductions you cannot substantiate, and you have run the whole year blind to your own margins.
• A website that does not say what you sell or how to buy. Cheap is fine. Unclear is not. Most bad small business sites are not ugly, they are vague.
• Not registering the entity. State filing is $50 – $500. Operating as a sole proprietor because the paperwork felt like a hassle means your personal assets are the business assets.
• The cheapest possible product sample. In ecommerce, a returns rate driven by quality problems is the most expensive line item there is, and it shows up after the money is already spent.
• Doing your own taxes with inventory, payroll or multiple states. A $900 return preparation fee routinely saves several times that.

Runway is what actually kills the business

Setup cost gets you open. Runway keeps you open. Most failed businesses were not badly built — they simply ran out of months.

Then add the number people leave out entirely: your personal living costs. If your household needs $4,500 a month and you want six months of runway, that is $27,000 that has nothing to do with the business but will absolutely determine whether the business survives. Founders who budget the business meticulously and their own rent not at all are the most common version of this mistake.

So the real total is:

Setup cost + (fixed monthly business cost × months of runway) + (personal monthly cost × months of runway)

For a typical local service business: $12,000 setup + ($1,200 × 6) + ($4,000 × 6) = $43,200. That number is much larger than the $12,000 people quote, and it is the honest one. If it is out of reach, the answer is usually to start the business alongside a job and shorten the runway requirement to near zero — not to pretend the number is smaller.

Discovery money versus scale money

This is the distinction that matters more than any individual figure.

Discovery money is spent to answer a question. Will anyone pay for this? At what price? Through which channel? It should be small, time-boxed, and attached to a written question with a threshold that would change your mind. A landing page and $800 of ads to see whether twenty people will pre-order is discovery money. So is a $1,500 blueprint that tells you the model does not work before you build it.

Scale money is spent against a known ratio. You know a customer costs $60 to acquire and is worth $400, so more spend is arithmetic rather than a gamble. You know each unit sells through in six weeks, so a bigger order is a working-capital decision rather than a bet.

Almost every expensive failure comes from spending scale money on a discovery question. The $45,000 software build before one paying customer. The $20,000 inventory order before one sale. The full rebrand before anyone knows the name. In each case the money was not wasted because it bought bad work — it was wasted because it was spent on a question that a few hundred dollars would have answered.

Run the cheap experiment first, every time. Getting from an idea to a first paying customer is almost always cheaper than founders expect, and finding out is what buys you permission to spend the large numbers.

A sane order to spend the first $10,000

Roughly, and for most business types:

1. $0 – $500 — Answer the demand question. Conversations, a single page, a small ad test. If nobody bites, you have saved $9,500.
2. $300 – $900 — Become a real business. Entity, EIN, bank account, basic bookkeeping setup.
3. $600 – $2,000 — Insurance. Before delivery starts, not after the first incident.
4. $500 – $3,000 — A clear place to buy. Not a beautiful site. A clear one, with a working way to pay you.
5. $1,000 – $4,000 — The minimum viable delivery capability. Tools, first inventory run, first month of contractor capacity.
6. $500 – $2,000 — One acquisition channel, tested properly. One channel, enough budget to read the result, tracked well enough to know what happened.
7. Whatever remains — hold it. Cash in the account is what lets you survive a slow month and act fast on a good one.

Notice the sequence buys certainty before capability, and capability before appearance. Most founders reverse it. If you want a second opinion on where your own money should go first, that sequencing is most of what a blueprint engagement produces.

The short version

• Setup cost is driven by three things: whether you need inventory or equipment, whether the product must be built before it can be sold, and how much you do yourself. Everything else is detail.
• Realistic setup ranges: professional service $3,000 – $8,000, productized service $5,000 – $15,000, local service $8,000 – $20,000, ecommerce $15,000 – $40,000, SaaS $25,000 – $75,000 unless you write the code.
• The forgotten costs are insurance, roughly 3 percent to payment processors, $200 – $500 a month of software, bookkeeping and tax prep, and multi-state sales tax once you cross the thresholds.
• Budget setup cost plus six to twelve months of both business and personal fixed costs. That total is usually three to four times the number people quote, and it is the number that decides survival.
• Defer anything whose value depends on scale. Never defer insurance, a separate bank account, or clarity about what you sell.
• Keep discovery money small and attached to a question. Only spend scale money once you know a ratio.