How Much Does It Cost to Start a Business? 406 Real Starting Points
There is no single number, and anyone quoting an average is guessing. Across the 406 businesses we index in 22 countries, 246 are one-person operations …
There is no single number, and anyone quoting an average is guessing. Across the 406 businesses we index in 22 countries, 246 are one-person operations — meaning the dominant startup cost was the founder's own time — and 106 of the 406 disclose a clean monthly revenue figure, median $27K/mo, ranging from $6/mo to $2.2M/mo.
That range is what you should budget against. Not a national average, not an agency quote, but the spread of what real businesses earned and how thin the teams behind them actually were.
Table of contents
What does it really cost to start a business?
Start with the structural fact: 246 of our 406 indexed businesses are run by one person. When headcount is one, the startup cost is time plus fees, and the invoice-shaped estimates you have been reading describe a category most of these founders never entered.
Against that, 106 businesses disclose a clean monthly figure with a median of $27K/mo. Compare any budget you are considering to that median before you commit, because it tells you what the middle of a disclosing sample actually produces.
We will also be honest about what we do not have. Very few founders publish their startup spend, so cost content across the internet is largely invented averages presented with a straight face. We would rather show you 106 revenue disclosures and 38 documented failures than fabricate a dollar figure.
The full set is browsable in all indexed ideas, and the grading behind every entry is documented in how we grade the evidence.
What the real cost line items are
Four lines recur regardless of what you start: platform or marketplace fees, tooling and infrastructure, paid acquisition, and your own time. Only one of them is genuinely fixed and published; the other three are set by decisions you make weekly. Build the sheet before the product.
| Cost line | One-off or recurring | Who sets it | Disclosed in our index? |
|---|---|---|---|
| Platform / marketplace fees | Recurring, per transaction | The platform's published schedule | No — read the official source |
| Tooling and infrastructure | Recurring | Vendors | Rarely |
| Paid acquisition | Recurring, scales with growth | The ad auction | Almost never cleanly |
| Founder or team time | Recurring | You | Indirectly — 246 of 406 are solo |
Use official sources for the fixed lines rather than a blog's round number, and lay the whole thing out in the SBA's startup cost worksheet before you spend anything. Marketplace-specific fees vary enough that we cover them separately — see Etsy Shop Setup Fee.
The line that kills budgets is acquisition. It is unbounded, it recurs, and it is the one no startup-cost calculator asks about.
What builders actually spent, by name
Named businesses, exact disclosed figures, evidence class, and what their cost base looks like. Where a cost is undisclosed we say so rather than estimating it.
- ·[The Viral App Monetization Machine (100 Apps Dissected)](/projects/paywall-machine) — Cal AI and Lerna at $2M/mo each; LazyFit, CoinSnap and Impulse at $700K/mo each. Third-party verified. Small teams throughout. Startup spend: not disclosed.
- ·[App Portfolio Studio Model](/projects/app-studio) — peak $2.2M/mo across 15 apps, founder-reported. The cost base is visible in the structure: a marketing operator, a technical co-founder and a developer community.
- ·Cluely — $500K/mo (roughly $6M ARR), founder-reported. Two founders, about 13–14 people, plus 60 contracted creators. That contractor line is a real recurring cost, and it is a distribution cost, not a build cost.
- ·Outrank — pushing toward $1M/mo, founder-reported. A tiny team plus contractors, run from France. Gravl — $440K/mo with 70,000+ subscribers, third-party verified, grown from three people to 13–14.
- ·[Retake AI Selfie Editor](/projects/retake-ai-selfie-editor) — roughly $2M/mo, creator-relayed from third-party Sensor Tower data. Team assumed, not disclosed. Costs unknown.
One more that shows cost honestly: a solo creator cloning a Floe-style agency SaaS reported $1.3M/mo for the original benchmark while earning $0 himself (creator-relayed). Same build, no distribution, no revenue.
How much is each of these numbers actually worth?
The index splits 57 third-party verified, 184 founder-reported, 121 creator-relayed, 44 unproven — 59% verified or founder-reported. So four out of ten entries in the index rest on secondhand telling or nothing you can check, and we label every one of them.
| Evidence class | Count | What it means |
|---|---|---|
| Third-party verified | 57 | Confirmed outside the operator's own telling |
| Founder-reported | 184 | The operator's own claim |
| Creator-relayed | 121 | Repeated by someone who did not run it |
| Unproven | 44 | Stated, with nothing behind it |
Gravl's $440K/mo is verified. Quitter's roughly $500K/mo is creator-relayed, secondhand from an interview. Those are different claims, and the near-universal habit of listing them together unlabelled is precisely why budgeting from internet numbers goes wrong.
The unproven tier is not theoretical. Cleo claimed $60K MRR in 53 days with zero proof, while the same team's earlier product, Mentions, did $20K MRR in month one — and the video making the claim was itself part of the launch funnel. That is what an unlabelled number can be.
One caveat, stated once: this index over-represents businesses that survived and then chose to publish. Treat the $27K/mo median as a tool for judging other people's claims, not as a forecast for yours.
The cost that decides the outcome is not the build
Building is the cheap, bounded part. The median disclosing business here earns $27K/mo, and what separates it from the ones earning nothing is distribution — the same product with no channel earns $0, which the Floe clone demonstrated at $0 against a $1.3M/mo benchmark.
Leftclick is the counter-example that proves the same point: approximately $400K/month across businesses (founder-reported), with 1 Second Copy peaking at $92K/month, Leftclick scaling to $72K in a month, and $100K+ in consulting. Those are distribution businesses wearing different product costumes.
Demand can even precede spend entirely. Chai could not afford its GPT-3 bill, posted to crowdfund it, and got 80,000 sign-ups in a day that crashed the servers — later $30M/yr with 21 people. No budget line created that.
E-commerce founders should size the market they are entering from primary data — the U.S. Census quarterly e-commerce sales release is the source to use — and read our category-level breakdowns in the What It Really Pays hub.
How to use this without fooling yourself
Four checks for this week. Each traces to a figure above.
- 1.Fill in the SBA startup cost worksheet before choosing a product. Pair it with the SBA's guide to planning a business. If the acquisition line is blank, the plan is not finished.
- 2.Budget against $27K/mo, not against $2.2M/mo. The top of the range belongs to teams with contractors and portfolios. The median is the comparison tool.
- 3.Assume solo until proven otherwise. 246 of 406 indexed businesses are one person. If your plan requires hiring before revenue, ask what that hire buys that a channel would not.
- 4.Grade every number you budget from. Mark it verified, founder-reported, creator-relayed or unproven. With 44 unproven entries in the index, an unlabelled claim like Cleo's is exactly what you are trying to filter out — the same test applies in cost-heavy niches like How Much Does Self Publishing A Book Cost.
Frequently asked questions
Is $10,000 enough to start a business?
For most of what we index, yes — 246 of 406 businesses are one-person operations where the main input was time rather than capital. The disclosed revenue range runs from $6/mo to $2.2M/mo, so budget size does not predict the outcome. Distribution does.
Is it true that 90% of startups fail?
We do not publish a survival rate, because we do not have a clean sample to compute one. What we can say: 38 of our 406 indexed entries are documented failures, and our index over-represents businesses that survived and chose to disclose. Any failure rate drawn from success-heavy sources will understate the real one.
Is $5000 enough to start a business?
Often, given that 246 of the 406 businesses we index run on one person. The constraint is rarely the starting cash — it is the recurring acquisition cost, which the median disclosing business at $27K/mo has solved and the 38 documented failures generally have not.
How reliable are these how much does it cost to start a business figures?
Startup costs are mostly undisclosed, and we say so rather than inventing averages. The revenue figures we publish split 57 third-party verified, 184 founder-reported, 121 creator-relayed, 44 unproven — 59% verified or founder-reported. Every entry carries its class so you can discount accordingly.
How many of the 406 indexed businesses actually disclose a monthly number?
106, at a median of $27K/mo across a $6/mo to $2.2M/mo range. That is roughly one in four. The other three in four named a business without a checkable figure, which is exactly the gap most startup-cost content fills with guesswork.