A complete guide to going freelance, organised by money milestones

Two numbers describe the American freelance market, and they differ by a factor of six. The Bureau of Labor Statistics counted 11.9 million independent contractors, 7.4% of total employment, in a July 2023 reference week; it counts people whose main job is independent work, in a single week, and excludes side-gig freelancing entirely. MBO Partners counted 72.9 million independents in a survey fielded in April 2025; it counts anyone who did any independent work at all, full-time, part-time or occasional, in the previous year.

These are not competing estimates of the same thing, and the difference between them is the actual subject of this guide. Most people who freelance are somewhere on the ramp between doing paid work on the side and having a business that is their main job. The ramp is where the money problems live, and it is almost entirely absent from guides that tell you to choose a niche, build a portfolio, set up a website and start pitching.

That sequencing is not wrong so much as beside the point. Nobody's freelance career ends because the website was late. It ends because a tax bill arrived that was larger than the bank balance, or because three months of invoiced work turned into two months of collected cash, or because a client stopped answering and there was no contract that said what happened next. Those are thresholds, not steps. You cross them in roughly the order below, and each one changes what your business is exposed to.

The milestones are not steps, and they don't arrive when you plan for them

The first five arrive in sequence for most people. The last two are triggered by conditions rather than reached by progress: you need a tax reserve system the month your liability outgrows a single month's income, whenever that happens, and you need a real contract the first time an engagement is big enough that non-payment would actually hurt. Either can arrive in month two or month twenty.

The rest of this post takes each row in turn. Every milestone here has a post of its own elsewhere in this series that goes deeper; the job of this one is to tell you which threshold you are standing at and what it changes.

Milestone zero: your runway is smaller than your bank balance

Runway is the number of months your cash covers your costs with no income at all. Two mistakes make the published version of this number wrong for almost everyone.

The first is measuring against your old salary. If you earned €4,200 a month and your fixed personal costs are €2,400, runway is calculated on the €2,400, plus whatever the business costs — call it €300 for software, accounting and insurance. At €2,700 a month, €16,200 of cash is six months of runway, not four.

The second mistake is more expensive. If you left employment part-way through a tax year, or you did paid freelance work before you went full-time, part of that balance is already owed. Tax reserve held inside your runway figure is not runway. Someone with €16,200 in the account and €3,800 of unpaid tax on last year's earnings has €12,400, which is four and a half months, not six. The runway calculator takes cash on hand, average monthly costs and expected income and returns the honest figure; subtract the tax before you enter the cash.

How many months you need is a separate argument, and it belongs with the other pre-departure tests: the three financial tests worth running before you hand in notice covers runway, demand evidence and the replacement cost of employer benefits. What matters here is that the number you calculate is the honest one.

The first paying client proves less than it feels like it proves

The first invoice being paid is a real event, and it changes two things that are easy to miss.

It creates a filing obligation. In the US, you must file a return once net earnings from self-employment reach $400, which is a low bar deliberately. In the UK there is a £1,000 trading allowance of gross trading income per tax year; below that you don't need to tell HMRC at all, and above it you either deduct the £1,000 or your actual expenses, never both.

It also creates the first data point you will ever have about your own pricing, and most people throw it away. The quoted rate is not the number that matters. Fee divided by every hour the project consumed (scoping calls, the revision you didn't charge for, the invoice you chased twice) is the number that matters, and you can only calculate it if you recorded the hours while doing them. Run your first finished project through the effective hourly rate calculator and treat the result as your actual rate. The full argument, including why published rate averages don't transfer, is in the post on what your rate actually is.

One US-specific change makes recording non-negotiable now. The 1099-NEC reporting threshold rose from $600 to $2,000 for payments made after 31 December 2025, under Pub. L. 119-21 §70433. A client who pays you $1,500 will send you no form and still owes you nothing in paperwork. The income is exactly as taxable as it was. Your own record is now the only record of it.

The second project from the same client is the first real signal

One client is a transaction. The same client coming back is the first evidence that what you sold was worth buying twice, and it is also the cheapest revenue you will ever earn, because the selling has already happened.

What breaks here is scope, and it breaks in a specific way: the second project gets quoted from your memory of the first, and your memory of the first is wrong. Suppose the first job was €3,000 against a plan of 30 hours, so €100 an hour. It actually took 46 hours once you count the two rounds of revisions and the kickoff call, which is €65.22 an hour. Quote the second at €3,000 and you have not repeated a €100 project. You have institutionalized a €65 one.

The fix is unglamorous. Before the second engagement starts, write down what changed from the first, whether that is more stakeholders, a tighter deadline or an extra review round, and price the difference explicitly. The questions that predict whether a project will be profitable is the checklist version of this. If the relationship is heading toward continuous work rather than discrete projects, the retainer structures that genuinely stabilize cash is the next thing to read, because only some of them do.

The month your revenue covers fixed costs, and the month your cash does

These are not the same month, and the distance between them is where most first-year cash crises sit.

Revenue covering costs means you invoiced enough. Cash covering costs means the money arrived. The gap is time-to-cash: the days between finishing work and having spendable money. It has four components, and only one of them is the client being slow — the others are how long you took to invoice, whether the invoice went to a person who can pay it, and how the payment rail settles.

The scale of the client-side portion is measurable. In IPSE's late payment research, 35% of UK self-employed respondents were not paid on time by a client in the preceding twelve months, and the average amount owed was £5,230. For a freelancer whose fixed costs are €2,700 a month, an overdue balance of that size is roughly two months of costs sitting in someone else's account.

So the milestone to actually track is collected cash against fixed costs, not invoiced revenue against fixed costs. Measure the delay per client rather than as an average, because the average hides the one client who takes 60 days. Time-to-cash and where freelancers lose money breaks the interval into its components, and the guide to getting paid on time covers the deposit structure, invoice timing and reminder schedule that compress it.

Paying yourself the same amount twice is a variance problem, not an income problem

At this milestone the business is viable and it still feels precarious, which confuses people. The reason is that your income has a mean and your rent doesn't.

Take three consecutive months of collected cash at €6,000, €1,800 and €4,400. The average is €4,067, and the average is useless to you: in month two you can pay yourself €1,800. Anyone drawing €4,067 across those three months is funding month two from savings and calling it income. The fix is to set your draw at a level the worst month supports and hold the surplus in a separate balance, so that months one and three fund month two rather than your personal account absorbing the swing.

This is also the point at which a forward view starts earning its keep, because a bank balance tells you about the past. A rolling thirteen-week projection of expected receipts against known costs tells you which week goes negative, which is the only thing you can actually act on. Worklyn's 13 Weeks Out does this from invoices and bank feeds rather than from a spreadsheet you have to maintain. When the forecast shows a gap that is structural rather than seasonal, planning for a slow quarter before it arrives covers the three levers in order of how fast they work.

Do not read a quiet quarter as a personal failure without checking the market first. Fiverr's Q2 2026 results reported annual active buyers down 21.9% year over year while annual spend per buyer rose 15.6% to $368. Fewer buyers, each spending more. If your work sits at the commodity end of a market doing that, the volume problem is not about you, and the answer is positioning rather than effort. Where freelance work actually comes from in 2026 ranks the channels by hours invested per client won.

The tax reserve becomes a system the month it outgrows your income

There is no date for this milestone. The trigger is arithmetic: the moment the amount you owe exceeds what you could comfortably pay out of a single month's profit, holding it in your head stops working.

In the US, self-employment tax is 15.3% applied to 92.35% of net profit, which is an effective 14.13% of profit before any income tax at all. On $60,000 of net profit that is $60,000 × 0.9235 = $55,410, and $55,410 × 0.153 = $8,478. Income tax sits on top. And it is not payable once a year: the 2026 Form 1040-ES requires estimated payments if you expect to owe $1,000 or more after withholding, on 15 April, 15 June and 15 September 2026 and 15 January 2027.

In the UK the marginal position between the Class 4 Lower Profits Limit and the Upper Profits Limit is 20% income tax plus 6% Class 4 National Insurance, so 26% on profits between £12,570 and £50,270. The part that surprises people is payments on account. Your first Self Assessment bill, due 31 January 2027 for the 2025/26 year, is the balancing payment for that year plus a payment on account of 50% of it for the next, with the second half due 31 July 2027. In cash terms the January bill is 150% of a normal year's tax, and it lands in the quietest month of most freelancers' year.

There is a second UK change that arrived this year and that a lot of people have not registered. Making Tax Digital for Income Tax went live on 6 April 2026 for anyone whose qualifying income was above £50,000 in 2024/25, and qualifying income is gross self-employment and property income combined, before expenses. It brings quarterly updates with their own deadlines. The threshold drops to £30,000 from April 2027 and £20,000 from April 2028, so most working freelancers are inside it within two years.

Practically: set aside a percentage of every payment on the day it lands, into an account you do not spend from. The tax set-aside calculator will give you a monthly figure from expected revenue, expenses and a country preset for the US, UK or Germany. How much freelancers should set aside for taxes works through the percentage, and why tax season is a year-round system covers the record-keeping side and the MTD deadlines in full.

One more thing belongs here, because it is a milestone people only notice in retrospect: your records have to outlive your tools. Fiverr Workspace closed on 1 March 2026, and Midday announced on 7 May 2026 that it was joining Ramp and winding down with 90 days of account access for export. Both gave notice, both gave an export path, and in both cases the people who came out fine were the ones whose invoices, contracts and hours were somewhere they could take with them. What the money side looks like once Midday is gone is set out on the Midday alternative page, and if you are moving off it, the Midday migration path imports from the export zip or an API token. If you were on Fiverr Workspace, where that work moves next covers the same ground for a saved export.

The contract milestone is triggered, not reached

You need a real contract the first time one of three things is true: the engagement is large enough that not being paid would materially hurt, the client has a procurement process, or the client sends you their paper to sign.

What changes at this point is that your default legal position becomes worth protecting. In the EU, Directive 2011/7/EU sets a default B2B payment period of 30 calendar days from receipt of the invoice, caps B2B terms at 60 days unless expressly agreed and not grossly unfair, and entitles you to statutory interest at the ECB reference rate plus at least eight percentage points, plus at least €40 per late invoice in fixed recovery compensation. With the ECB main refinancing rate at 2.40% since 17 June 2026, the minimum statutory rate for the second half of 2026 is 10.40%, and several Member States apply more. In the UK, the Late Payment of Commercial Debts (Interest) Act 1998 gives Bank of England base rate plus 8%, which with Bank Rate at 3.75% means 11.75% for 1 July to 31 December 2026, plus a fixed sum of £40, £70 or £100 depending on the size of the debt.

Work one through. A €4,000 invoice paid 45 days late at 10.40% earns €4,000 × 0.104 = €416 a year, or €1.14 a day, so €51.29 over 45 days, plus the €40 fixed compensation: €91.29. That is not life-changing money, and it is not the point. The point is that these entitlements exist automatically, and a contract can weaken them. A clause setting 90-day terms, or one that excludes interest, is exactly what a badly-drafted agreement gives away without you noticing you agreed to it.

So the contract you need is not elaborate. It needs the scope boundary, what happens on a change, the payment terms and the position on late payment, a suspension right, and who owns the work before it is paid for. The contract generator produces a signable PDF in the browser with no signup, and Worklyn's contracts feature adds e-signature and a clause reader that flags the terms worth arguing about. When it does go wrong, the reminder schedule that works without souring the relationship is the first response, not the letter.

Where the threshold actually sits

The milestone that decides whether you are still on the ramp is the fourth one, and it has a formula.

Let F be your monthly fixed costs, personal and business combined, and r your tax reserve rate. The collected cash you need in a month is not F. It is F ÷ (1 − r). At €2,700 of fixed costs and a 30% reserve, that is €2,700 ÷ 0.7 = €3,857 collected. Not €3,857 invoiced, and not €2,700.

The rule: when collected cash clears F ÷ (1 − r) for three consecutive months, you are past the ramp, and the binding constraint on your income changes from volume to price. Below that line, every hour you spend should go into shortening time-to-cash and winning repeat work, because more clients at the same collection speed just makes the gap bigger. Above it, more clients is the wrong lever, and choosing the pricing model that fits the work is the right one. That is also the point at which the operating habits of freelancers who last start to matter more than any single decision, and the point at which Worklyn for freelancers is worth a look as one place for the proposals, contracts, hours and invoices that currently live in four.

Worklyn's Safe to Spend and 13 Weeks Out work from collected cash and known costs rather than from your bank balance, which is the difference between knowing what you have and knowing what is yours.