Statement of work vs proposal, and which one binds you.

A proposal is the document you send to win the work, and a statement of work is the document that defines the work once you've won it. The proposal sells an outcome and stays deliberately loose about the details, while the statement of work fixes the deliverables, the dates, the revision count, the exclusions and the fee. On most agency projects the proposal goes out first and the statement of work is the one that's supposed to govern. Clients quote the proposal anyway, because the proposal is where you were at your most persuasive.

Look this up and you'll mostly find it explained from the client's side, by contract platforms and procurement blogs written for the company buying the work rather than the agency selling it, and not one of them prints a sentence of either document or a single dollar figure. So here's the same comparison from the other side of the table, with one project written out twice: a nine week website build priced at $26,000 at a blended rate of $150 an hour. Every number below is arithmetic on that example rather than an industry average.

What's the difference between a statement of work and a proposal?

A proposal argues that you should get the job, and a statement of work defines the job you got. The proposal is written for somebody who hasn't decided yet, so ranges are fine and the tone is allowed to be warm, since the document's job is to win a decision.

A statement of work is written for somebody who's annoyed in month three. It names deliverables that can be counted rather than described, puts real dates against them, says how many revision rounds are included and what a round even means, prices every item that sits outside the fee, and gives each deliverable a condition that says when it's finished. Writing one properly is its own craft, and the section by section version of that job is worth reading before your next one goes out.

The test that settles most drafting arguments is short. Anything a competing agency could also claim about themselves belongs in the proposal, and anything you'd want to point at during a disagreement belongs in the statement of work. A client's procurement team will also ask for a scope of work when they mean a statement of work, and the difference between a scope of work and a statement of work trips agencies up more often than this one does.

Which one comes first, and does the proposal become part of the contract?

The proposal comes first, and it becomes part of the contract only if the contract says so, which is why most disputes over a proposal are really disputes about a missing sentence. The advice you'll read on the order contradicts itself, one version running proposal to contract to statement of work and another running request for proposal to proposal to a contract with the scope attached, and neither one acknowledges that the other exists.

For an agency the version that causes the least trouble is to win the work with the proposal, sign one agreement carrying the commercial terms, and attach the statement of work to that agreement. Sending the scope afterwards as a loose follow-up nobody countersigns leaves you with a signed proposal and an unsigned scope. Where a master services agreement already exists the scope is issued under it and inherits its terms, which is the arrangement covered in the MSA and SOW comparison.

What does the same deliverable look like in a proposal and in a statement of work?

The proposal describes the deliverable in a sentence the client will like, and the statement of work describes it in a sentence the client can't reinterpret. Here's the design and build line from that $26,000 project in each document.

Three changes do most of the work there, and each one closes a specific argument. "A set of flexible page templates" became eight named templates, so the client can count them instead of imagining them. "Rounds of feedback" became two rounds with a definition and a five day window, which stops a trickle of comments over three weeks from consuming the revision budget. And the ninth template, the one that always gets asked for, has a price and a delay next to it rather than a conversation. None of that makes the proposal dishonest, since it was written to be liked, but the two phrases doing the damage are the ones that read as most generous.

What happens when the proposal and the statement of work disagree?

Whichever document the agreement names as controlling wins, and if it names neither, both sides argue about intent instead of pointing at a sentence. Plenty of templates carry a generic entire agreement paragraph that supersedes all prior discussions without ever naming the proposal, and naming it by date is the part that does the work.

That last sentence is the one worth arguing for if your client's lawyer wants the clause trimmed. Without it, a client can accept that the scope controls on any point of conflict and still insist a deliverable mentioned only in the proposal was never in conflict with anything, so it stays in.

When the client signs only the proposal, that proposal becomes your scope, and the way out is to convert it rather than replace it: a single page addendum adds the counts, the dates, the exclusions and the change path on top of what they already agreed to. The longer version of that problem, including what a signed proposal actually gets you, is worked through separately.

The alternative is to accept that some clients will never sign a second document, and to write proposals that could survive being the only one. The deliverables section loses its unmeasurable adjectives, gains counts, and carries a short priced exclusions paragraph even though it sits inside a sales document.

Is a statement of work legally binding and a proposal not?

Not reliably, in either direction, so the confident version of that claim is worth ignoring wherever you run into it. A proposal that identifies both parties, describes the work, states a price and gets accepted in writing can function as an enforceable contract in many places, while a statement of work sitting on its own with nobody's name at the bottom is closer to a description of work. What creates the obligation is the agreement and the acceptance rather than the label at the top of the page.

Two caveats belong here before you paste any of that into a live document. Nobody here is a lawyer, so the clauses above are drafting starting points, and whether a given sentence binds anybody turns on your jurisdiction and on facts we can't see from here. The second is narrower: what Docket records is which version of the scope both sides approved and when, which is a different job from proving anything to a court.

Where does a request for proposal fit in?

The request for proposal (RFP) comes first, your proposal answers it, and the statement of work comes last, after you've won. A request for quote (RFQ) sits in the same slot as an RFP and asks a narrower question: an RFP asks how you would solve the problem and what that costs, while an RFQ asks only for a price against a spec the client has already written, which is the difference between quoting and proposing seen from the buying side.

The trap after you win one is specific, because an agency that wins a request for proposal often copies the client's requirement list straight into the scope, on the logic that the client wrote it so the client must agree with it. But a wish list written before anybody costed it usually contains deliverables you never priced, so read every pasted line against your own estimate and move anything unpriced into the exclusions with a figure next to it.

What does one vague proposal line actually cost?

On this $26,000 project, three unpriced additions of roughly ten hours each come to 30 hours, or $4,500 at the blended rate of $150. Price it at a thirty percent margin and you had $7,800 of profit, so those three conversations took about 58 percent of it. They came from the proposal rather than from a difficult client, since "flexible page templates" invited the ninth one and "rounds of feedback" invited the third round, which is why what scope creep costs an agency across a year is mostly a documentation problem rather than a discipline one.

A proposal written to be liked isn't dishonest, but the phrases that read as most generous are the ones that end up costing you money.

Write the proposal as if it's the only document that gets signed

Start by adding one sentence to your proposal template, directly under the price, that points forward to the scope document you'll send next: "This investment is based on the scope described above. The final deliverable list, revision counts, exclusions and dates are set in the statement of work, which we'll send for signature within five business days of your go-ahead." Then delete every adjective in the deliverables section you couldn't check, and replace each one with a count, a name or a number. If you'd rather start from something with the counts already in place, a filled-in statement of work with real fees saves you the blank page.

Last, put the entire agreement clause into your scope template with a blank for the proposal date, and fill that blank every time. Then say it out loud at kickoff, because a client who hears that changes are welcome, priced within three days and approved by one named person takes it better in week one than in week seven. That's the argument behind telling the client at kickoff that changes get priced, and the habit that keeps it honest afterwards is a change request form with the impact already priced rather than an email thread.

Frequently asked questions