Proposal vs contract, and which one governs the work.

A proposal is written to win the work and a contract is written to run it, so the two documents describe the same project at completely different levels of precision. The proposal makes the case and gives a number. The contract fixes the countable version of the work, the payment schedule, the revision limit and the route for changing any of it. Both usually get signed, which leaves you with two descriptions of one project, and the question nobody answers until it costs money is which of them governs when they disagree.

Most explanations stop at the split between selling and protecting, tell you to send both documents, and move on. That's where the real exposure begins, because the proposal doesn't stop existing once the contract is signed. Your client read it more carefully than anything you sent afterwards, and its friendliest sentence is the one they'll quote back at you. What follows is the same project written both ways, the arithmetic on the gap between them, and the clause that settles it in advance.

What's the difference between a proposal and a contract?

The proposal is an argument and the contract is a set of limits, and almost every practical difference between them comes out of that. A proposal is written for a reader who hasn't decided yet, so it explains the approach, shows relevant work, gives a fee and a rough timeline, and stays deliberately warm about the edges. A contract is written for a reader who has already said yes, so it counts things: how many pages, how many rounds, by when, for how much, and what happens when the answer to any of those changes.

Neither document is trying to do the other's job. A proposal stuffed with liability clauses and precise revision limits converts worse, because it reads as though you're bracing for a fight before the work has started. A phrase like "revisions until the direction feels right" is lovely in a pitch and unusable as a boundary once someone is on round four.

The split shows up most clearly in how each one treats detail. A proposal happily says "annual report layout" because the client hasn't sent the content yet and nobody knows how long the report will be. The contract has to name a number anyway, since the difference between a 32-page report and a 44-page one is real work somebody has to pay for. The same tension runs through the statement of work and the proposal, and it's why those two documents sit on either side of the yes.

Which document governs when the two say different things?

Whichever one the contract names, and if the contract names neither then you're in an argument rather than a lookup. This is the whole ballgame, and it takes about three lines to settle.

Think about how the disagreement actually arrives. Nobody sits down to compare the two documents line by line. In week five a request comes in, you say it's extra, the client goes looking for the sentence that says otherwise, and they find it in the proposal because that's the document they read closely enough to approve spending money on. Your contract may be far more precise, but precision only helps if the contract also says it outranks everything that came before it.

Two clauses do this work together, and most agency contracts carry a weak version of one and none of the other. The entire-agreement clause says this document replaces prior discussions, which is common enough that people paste it in without thinking. The order-of-precedence clause ranks the documents that survive, which is what you actually need when the proposal has been attached as an exhibit rather than thrown away. Without the ranking, an attached proposal has the same standing as the scope document sitting next to it.

What does the same project look like in the proposal and in the contract?

It looks like the same five deliverables described in two vocabularies, one of which has numbers in it. Here's a seven-week annual report and investor deck project for a financial advisory firm, written first the way it went out to win the work and then the way it appeared in the signed agreement.

The second version isn't more honest than the first, it's just countable. Every rate in it comes out of the fee rather than being invented for the occasion, since $6,400 across 32 pages is $200 a page and $3,600 across 18 slides is the same $200 a slide. A client can do that division as easily as you can, which is exactly why it's worth showing rather than hiding.

The proposal you sent in January is still in the room in week five, and your client remembers it better than you do.

Now watch what the gap between the two versions is worth. Calder & Finch send their approved copy in week four and the report runs to 44 pages, which is 12 pages past the number in the agreement at $200 each, so $2,400. They also want a third pass on the deck after the second round closes, which took six hours at $180, so $1,080. Together that's $3,480, or just over a fifth of the whole fee, and every dollar of it turns on which document the answer comes from. Under the agreement it's a priced change. Under the proposal's "annual report layout" and "revisions until the direction feels right", it's work you already promised.

What does a signed proposal actually get you?

It gets you whatever the proposal happens to say, which is the part agencies underestimate in both directions. A signed proposal that names the work, the fee and the payment terms isn't nothing, and clients who assume they can walk away from one are often wrong. It's also not the protection people imagine, because a signature doesn't add the detail the document never had.

Go back to Calder & Finch. If the only signed paper in the file is the January proposal, you have written agreement on $16,400 and on five deliverables named in five short phrases. What you don't have is a page count, a slide count, a chart count, a revision limit, an hourly rate for anything, a deadline for the client's own copy, or a defined route for changing any of it. So the $3,480 conversation in week five has nothing to rest on except which of you sounds more reasonable, and the client is the one holding a document that promised revisions until the direction feels right.

Whether a particular signed proposal binds a particular client is a question for a lawyer in your state. The version you can act on this week is narrower: whatever it binds them to is the text you wrote while you were trying to be likeable, which is a poor place to keep your only description of the work. That's the reason the statement of work and the contract are worth keeping as separate documents that reference each other.

How do you word the clause that ranks the documents?

You name the documents, put them in order, and say what happens where they conflict. Three sentences, one section, and it belongs near the back of the agreement with the other general terms.

The third sentence is the one worth arguing for, and it's missing from almost every template version of this clause. Ranking the proposal last protects you on page counts and revision rounds. Keeping it alive where the scope document says nothing protects the client on the things they actually bought, which is the approach, the thinking and the promise that one art director would run the whole suite rather than three freelancers. Deleting the proposal outright would strip that out, and a client who notices is right to push back.

A second sentence belongs on the other side of the file, in the proposal itself. A single line saying the proposal describes the intended approach and that a signed agreement and scope document will define the deliverables and terms costs you nothing in a pitch. Both lines go into your templates once and neither has to be renegotiated per project.

What happens to a change request when only the proposal was signed?

It stops being a change request and becomes a favor you're asking for, which is a much weaker position than it sounds. A change request works because both sides already agreed what the baseline was, so the only open question is the price of moving off it.

With a countable scope document in place, the 44-page report is a two-line message: the agreement covers 32 pages, the extra 12 are $200 each, please confirm the $2,400 and the two extra days before we lay them out. Nobody has to be persuaded of anything, because the client approved the page count and the rate in February and is simply being shown the arithmetic. With only the proposal in place, the same message has to first establish that 32 pages was ever the deal, and you're introducing a limit that appears nowhere in writing at the exact moment it costs your client money.

This is why the change route belongs in the agreement rather than in an email habit. Name who can approve a change on each side, say that changes take effect only in writing, and give yourself a standing change request form so raising one is cheap enough that it always happens. The requests that quietly turn into free work are almost never refused, they're just never priced, and the same pattern shows up across most in-scope and out-of-scope disputes on fixed-fee projects.

Do you need both documents on every project?

You need both whenever the work gets priced before it's fully defined, which is most agency work and all of the fixed-fee kind. The proposal exists because the client has to choose you before anyone can afford to specify the job properly, and the contract exists because that specification has to happen anyway before somebody starts laying out pages.

The exception is the repeat client you already have a signed agreement with. Once the terms, the rates, the payment schedule and the change route are settled from the first engagement, the second project needs a scope document referencing that agreement and nothing more, which is the arrangement a master service agreement and a statement of work are built to give you. That's a real saving on a client relationship that runs for years, and it's worth setting up on the second project rather than the fifth.

What doesn't work is the collapsed version, where the proposal grows a terms section and gets treated as the whole file. The terms end up too generic to settle anything specific while the deliverables stay in pitch language, so the document is harder to read and no more useful in week five. The countable version has to exist somewhere, and a scope of work clients can't misread is where it belongs.

Why a proposal only costs you money after it wins

Go back to the two Calder & Finch documents and the line that decides the project isn't the $16,400 either. It's "annual report layout" against "annual report layout, up to 32 pages", and the four words at the end of the second one are worth $2,400 the moment the client's copy runs long.

That trade runs one way for as long as the proposal is out for review, when its vague lines keep the reader focused on whether you're the right agency rather than on page counts nobody could have known in January. The day it's accepted, those same lines flip from an asset to a liability, and they stay in the file for the length of the project unless something outranks them. Losing proposals cost you a week of writing, and winning ones cost you whatever the difference is between what you meant and what you wrote.

Docket picks up from the point where the request arrives. The 12 extra pages, the $2,400 and the written yes from the person named in the agreement end up on one record attached to the project, rather than spread across a month of email nobody can find in April. That doesn't settle which document governs, since the clause above is what does that, but once the answer is agreed there's a single place showing what was asked, what it cost and who approved it.

Frequently asked questions