Almost all advice about freelance contracts is advice about the rate. The rate is the easiest thing to discuss, it is the number both sides showed up to talk about, and it is rarely the term that wrecks a project. The terms that do the damage are the boring ones: what counts as finished, who decides, when money is owed, and what happens when someone wants out.
Freelancing is not a fringe activity. Upwork's Freelance Forward research estimated that 59 million Americans had done freelance work in the preceding year. Treat that as an order of magnitude rather than a measurement — it is a commissioned survey from a company whose business depends on the number being large, and definitions of "freelance work" stretch from full-time consultancy to one paid weekend. The point stands anyway: a very large number of people are signing agreements, and many of them are signing without reading past the fee.
I should say plainly where this comes from. I am not a freelancer. I am an engineer who leads research teams, and I have never made a living from client work, so nothing here is dressed up as personal war stories. It is written from the structure of these agreements and from what the standard clauses actually do.
And it is not legal advice. That sentence is usually a reflex; here it is load-bearing. Contract law is jurisdictional in ways that matter enormously. A clause that is routine and enforceable in California may be unenforceable, or mean something entirely different, in India or the EU. What follows describes what common clauses do and which ones deserve suspicion. For anything substantial — a long engagement, a client far larger than you, work involving IP you care about, or any liability you could not absorb — paying a lawyer to read it once is cheap against what a bad clause costs.
Scope is the term that decides everything else
Scope creep is usually described as a client behaviour problem. More often it is a drafting problem. If the agreement says you will "design a website" or "build the data pipeline," then every disagreement about whether something is extra is a disagreement about the meaning of an English phrase, and the party with more money and more patience tends to win those.
A scope section that works is boring and specific. It names the deliverables and their formats. It names the number of rounds of revision included, and what a round is — one consolidated set of feedback, not seventeen messages over three weeks. It names what is explicitly not included, which is the part people skip and the part that does the most work. And it states what the client owes you: feedback within a defined window, access to systems, a single named person who can approve things.
That last one is underrated. A project with three approvers and no tiebreaker will consume more of your life than one with a difficult client and a clear decision-maker.
Acceptance: what counts as finished
Somewhere in the contract there is a moment when the work is deemed accepted, and payment usually hangs off it. Read that clause slowly.
The dangerous version makes acceptance depend on the client's satisfaction, judgement, or approval, without qualification. If the trigger for your final payment is somebody's subjective happiness, you have written them an option to withhold money indefinitely at no cost to themselves. The safer versions tie acceptance to objective criteria — the deliverables listed in the scope, delivered in the stated format, meeting specified tests — and add deemed acceptance: if the client does not reject in writing with specific reasons within a set number of days, the work is accepted.
Deemed-acceptance language is genuinely important for freelancers, because silence is the most common way projects stall. It does not make a client pay. It does make it much harder for them to argue later that they never accepted anything.
What the payment clause actually promises
Payment terms carry more of your financial risk than the rate does. Net 60 on a good rate is worse than net 15 on a mediocre one if you are the one financing the client's cash flow.
Things worth having: a deposit or first milestone paid before substantial work begins, payments tied to milestones rather than one lump at the end, and a clear statement of what triggers each payment — a date, a delivery, an acceptance event. Interest on late payment is common, though whether the stated rate is enforceable depends on where you are, and a late-fee clause mostly works as a signal rather than a remedy. Small freelancers almost never sue, and sophisticated clients know it. The deposit is the real protection.
Also look for a termination clause. Many contracts let the client terminate for convenience with short notice. That is not automatically unfair, but it should come with payment for work performed up to termination, and for a substantial engagement, some compensation for the booked time you can no longer fill. What you want to avoid is a contract that lets them stop on Tuesday and owe you nothing for the six weeks you held open.
Who owns what you made
IP is the clause most often handled by vague goodwill, and the one where jurisdiction bites hardest.
In the United States, "work made for hire" is a specific statutory category that only applies to certain kinds of commissioned work and only when agreed in a signed writing — clients use the phrase far more loosely than the law does. In India and the UK, copyright in commissioned work generally starts with the author unless there is a written assignment, and the assignment has to actually be in writing to transfer anything. Many jurisdictions also have moral rights that cannot be assigned at all, only waived, and some do not permit even that.
The practical questions to settle: does ownership transfer, or is the client getting a licence? If it transfers, does it transfer on final payment rather than on delivery? Can you show the work in your portfolio, and is there an embargo period? Do you keep the right to reuse your own generic tooling, templates and libraries — because if you sign away everything you touched, you have sold your future productivity along with the project.
The clauses that are quietly expensive
A few things show up in client-drafted agreements that are worth pausing over.
- Unlimited or uncapped liability. Your exposure should normally be capped at something like the fees paid under the contract. An uncapped indemnity means a single project can cost you far more than it ever paid.
- Broad indemnities. Indemnifying a client against claims arising from your own work is one thing. Indemnifying them against claims arising from their use of it, or from materials they supplied, is quite another.
- Exclusivity and non-competes. A clause preventing you from working with a whole industry for a year is a large, usually uncompensated restriction on a person whose income depends on having many clients. Enforceability varies a lot by country, but do not rely on unenforceability as a plan.
- Unlimited revisions at a fixed fee. This is not generosity, it is an unbounded liability with a friendly name.
- Confidentiality that runs forever and one way. Fine in principle; check that it does not also prevent you saying who the client was.
If you are freelancing from India
A few structural things differ enough to matter, and none of this replaces talking to a chartered accountant.
Clients deduct TDS on professional fees, which means the money landing in your account is less than the invoice and you reclaim the difference when you file. Plan cash flow around the net figure, not the gross. GST registration has a turnover threshold, and export of services to foreign clients can be zero-rated if the conditions are met, typically by furnishing a letter of undertaking — but "export" has a specific definition and getting it wrong is expensive. If you take foreign payments, understand the remittance path, the bank charges, the spread, and what documentation your bank wants.
On enforcement, be realistic. Cross-border contract disputes for amounts typical of freelance work are almost never economically worth litigating, whatever the jurisdiction clause says. This is precisely why deposits, milestones and staged delivery matter more than remedies. Your leverage is unfinished work in your possession, not a courtroom.
Negotiating without a script
The negotiation advice industry sells scripts that supposedly work. They do not reliably work, because the outcome depends mostly on something the script cannot change: whether the client has alternatives and whether you do.
If you are early in your career with a thin portfolio and no pipeline, your leverage is genuinely limited, and pretending otherwise mainly loses deals. If you have specific expertise the client cannot easily replace, or a queue of other work, you can hold a line. Knowing honestly which situation you are in is worth more than any phrasing.
What is generally true is that negotiating across several variables goes better than pushing on one. A client with a hard budget ceiling may still have room on payment schedule, timeline, revision rounds, portfolio rights, or the size of the first milestone. Asking which constraints are real is a reasonable question and usually gets an honest answer.
It is also true that asking has a real, if small, cost in some situations, and advice that insists there is never any downside is not being straight with you. A client who reacts badly to a polite, specific request about payment terms has told you something useful about the next six months. That is information, not a disaster — but it is not nothing either.
Before you sign
Read the whole thing, including the parts that look like boilerplate, because termination, liability and IP usually live there. Check that the scope in the contract matches the scope you discussed, since they drift. Make sure the payment triggers are events you control or can prove. And if a clause is written so that you cannot tell what it means, that is not your failure of comprehension — ambiguity in a contract benefits whoever is better resourced to argue about it later, and that is not you.
Tags
Taresh Sharan
support@sharaninitiatives.com