A proposal is not a document that describes a company. It is the document on which a named person decides whether to commit money, and everything else about it is decoration. Read that way, the work reorganises itself. If the reader is deciding, then the proposal’s first job is to let them decide quickly: every requirement they wrote answered in the place they wrote it, every term that qualifies the answer stated next to it, every figure traceable to a document, and one unambiguous request at the end that says what accepting looks like, who signs it, and by when.
Most proposals fail on none of these points and on all of them at once. They open with the seller: history, certifications, values, the strength of the team. They describe the product in the seller’s vocabulary. They price in a table nobody can reconcile to the brief. They close with “we look forward to hearing from you”, which is not a request, and which leaves the buyer to work out what happens next. The document is competent and it asks the reader to do the work of turning it into a decision.
This piece sets out the eight sections that do that work, in the order a decision is taken; the difference between a close and pressure; a worked example of a nine-requirement brief that produces an offer of $22,720.00 and a total of $24,960.00 once two items are settled; the case where the reader is a process rather than a person; and the limits of structure itself.
What the buyer is deciding
The decision behind a proposal is not “is this a good company”. It is a conjunction of four questions, and a proposal that answers three of them has not answered the decision.
Does it meet the requirement? The buyer wrote a brief, or said something in a meeting, or forwarded an email. Each of those contains requirements, some stated as facts and some as expectations. The proposal has to answer the ones that were written, one by one, in the buyer’s order — because the buyer will check in their order, not the seller’s.
At what total? Not at what unit price. The buyer’s total is the thing they will compare, and it includes freight, installation, tax treatment, the second year, and whatever the seller has excluded. A price that is right and a total that is wrong is a proposal that will be renegotiated at the worst possible moment.
On what terms? Payment days, warranty, liability, termination, validity. Terms are where a proposal stops being comparable: two offers at the same total on different payment terms are not the same offer, and the buyer’s comparison will normalise them whether the seller does or not.
By when, and from whom? A delivery date the seller can hold, and a counterparty the buyer can hold to it. That is where a legal entity, an insurance position and a named account contact become part of the commercial offer rather than an appendix.
A proposal answers those four and it is a decision document. It does not need to persuade anybody that the company is admirable, because the buyer did not ask; it needs to leave no question unanswered that the buyer would have to ask before deciding.
The eight sections
Table 1 is the structure. Each row names a section, the question the buyer brings to it, and the failure that follows when the section is missing. The right-hand column is the reason the structure is not a matter of taste: every missing section becomes a question the buyer has to ask, and a question is a delay at best and a reason to choose the other offer at worst.
Table 1. The eight sections of a proposal, the buyer’s question that each answers, and what fails when the section is absent. Source: the sales role’s tailored proposal, which restates the buyer’s requirements and carries every figure back to the document it came from.
| Section | The buyer’s question | What fails when it is missing |
|---|---|---|
| 1. The requirement, restated in the buyer’s words, with the brief’s date | Did they read what I sent? | The buyer re-reads their own brief to check; the proposal becomes a comparison of documents rather than offers |
| 2. The answer to each requirement, numbered to the brief | Does it meet it? | Requirements are answered out of order or not at all, and the buyer’s scoring sheet records gaps that do not exist |
| 3. The terms as supplied, with the qualifying clause quoted | On what conditions? | The buyer assumes standard terms and discovers the qualification after selection |
| 4. The price and its basis: unit, quantity, currency, inclusions, exclusions, validity | At what total? | The buyer’s total and the seller’s total differ, and the difference emerges in the contract |
| 5. Delivery: date, point, and what the seller needs from the buyer to hold it | By when? | The date is treated as a promise rather than a plan, and the first slip is a dispute |
| 6. The open questions, marked, with what would close each | What is not yet known? | The gap is filled by an assumption, and the assumption becomes a commitment when the buyer relies on it |
| 7. The comparison the buyer will run, field by field | Against what? | The buyer builds the comparison without the seller’s fields, and the offer is a row in somebody else’s table |
| 8. The decision requested: what acceptance is, who signs, by when, and what happens next | What do I do now? | The proposal joins a queue of documents that are all waiting for something |
Two rows carry more weight than their length suggests. Section 6 is the one sellers delete, because a marked gap looks like an admission. It is the opposite: a gap that is marked is a question the buyer can answer, and a gap that is filled with a plausible number is a commitment the seller has made without authority. The sales role is built so that no figure originates inside it — where the client’s material is silent, the draft carries the gap and the question, and the buyer sees a seller who knows the difference between what is known and what is assumed.
Section 7 is the one sellers underestimate. The buyer will build a comparison; that is what buyers do with three proposals. A seller who supplies the fields — the buyer’s named alternatives, the terms, the delivery dates, the total basis — has shaped the table the buyer will use. This is not a trick and it only works when the fields are the buyer’s, not the seller’s: a table whose columns are chosen to make the seller win is a brochure with a grid on it, and buyers recognise it in about four seconds.
The order is the argument
Numbering the answers to the brief is the cheapest improvement available to any proposal, and it is the one most often skipped. A buyer with a nine-point brief and three proposals does not read three documents; they read three answers to point 1, then three answers to point 2. A proposal that follows that order is read in full; one that follows the seller’s own narrative is read until the buyer finds point 1, and skimmed after.
The restatement in section 1 is part of the same discipline rather than a courtesy. Restating the requirement in the buyer’s words does three things at once: it proves the brief was read, it fixes the version of the requirement being answered — briefs get amended, and the date of the version matters — and it gives the buyer a sentence they can lift into their own internal paper. That last one is worth more than it looks. The person taking the proposal to an approver needs language that survives the journey, and a seller who supplies it has removed a step.
The section order is also diagnostic. Where a seller cannot write section 2 — the answer to each requirement — because the material does not exist, the proposal is not ready. That is a finding about the offer rather than the writing, and it is much cheaper to discover at draft stage than after the send.
The close is a request, not pressure
A close is a dated question addressed to a named person, with the route to accept stated and the consequence of each answer described truthfully. That is the whole of it, and it is worth separating carefully from the thing it is often confused with.
A request says: the offer holds until 5 November; acceptance is a countersigned copy of page 12 or a purchase order referencing this proposal; delivery begins ten weeks after acceptance, which puts the first day on site at 14 January; if the answer is no, a sentence saying so closes the file and I will not write again this quarter. Every clause is true, checkable, and reversible by the buyer.
Pressure says: this price expires on Friday, when it does not; we have two other clients considering this slot, when there are none; or a discount that is available today and will not be available tomorrow, when the seller’s own price list makes no such distinction. Pressure is a claim about the world that the seller cannot support, deployed to shorten the buyer’s decision. It sometimes works, and it costs the account, because the buyer discovers the claim was false at the moment they would otherwise have become a customer.
The test that separates them is whether the buyer can say no and remain a prospect. A request that can be refused, in writing, and answered with a follow-up in six months is a request. An offer where refusal is made to feel expensive is pressure, and the expense is being manufactured by the seller rather than discovered by the buyer.
There is a second test, and it is about dates. A close has a date because the offer has a date — a validity period, a delivery window, a price list that gets revised. Where the date comes from the seller’s own documents, it is a fact and it belongs in the close. Where the date exists only to create urgency, it is pressure with a calendar attached, and the buyer who asks “why Friday” is owed an answer that traces to something.
Worked example: a nine-requirement brief, one offer, two items outstanding
The example is constructed for this piece. Meridian Care Group issues a brief dated 6 October 2026 for 1,200 units a month, delivered to two sites, with installation, a 24-month warranty, thirty-day payment terms and a first delivery by 14 January 2027. The seller’s price list carries 1,200 units at $18.40, freight at $320.00 a site, a second-year warranty extension at $1.15 a unit, and a ten-week lead time from acceptance. The terms sheet gives a twelve-month warranty, thirty days’ payment and a validity of thirty days from the date of offer.
The proposal is dated 6 October 2026 and holds to 5 November 2026 — thirty days. Acceptance by that date puts ten weeks on site on 14 January 2027, which is the date the brief asks for. That correspondence is not luck: it is the first thing to check before the offer goes out, and where the lead time does not fit, the proposal’s job is to say so in section 5 rather than to promise the date. Table 2 sets out the offer as it would be sent, with the source of every line.
Table 2. The offer as sent, with the source of every figure and the two items that are not in the total. Constructed for this piece.
| Line | Source | Amount |
|---|---|---|
| 1,200 units a month at $18.40 | Price list, item 1180 | $22,080.00 |
| Freight to two sites, $320.00 each | Price list, item FR-2 | $640.00 |
| Total as sent | $22,720.00 | |
| 24-month warranty: second year at $1.15 a unit, 1,200 units | Price list, item W-24 | $1,380.00, priced as an option, not in the total |
| Installation at two sites | No source in the client’s material | $0.00, marked open, question raised |
The arithmetic is short and it has to be, because the buyer will redo it. 1,200 × $18.40 = $22,080.00; freight is 2 × $320.00 = $640.00; the total is $22,720.00. The warranty extension is 1,200 × $1.15 = $1,380.00 and is carried as an option with its own line, because the terms sheet supplies twelve months and the brief asks for twenty-four: the difference is a fact about the offer, and the buyer is entitled to see it rather than to infer it. Installation has no source, so it has no figure; the proposal carries a marked gap and a question, which is what the role does instead of estimating.
Table 3 is the check that has to be run before the offer leaves: all nine requirements of the brief, in the brief’s order, and where each one is answered.
Table 3. The nine requirements of the brief, how each is answered, and the section of the proposal that carries the answer.
| Requirement | Answered in the proposal | Where |
|---|---|---|
| 1. Two sites, five days a week | In the total, with both sites named | Sections 2 and 5 |
| 2. 1,200 units a month, monthly billing | In the total, as the quantity basis | Sections 2 and 4 |
| 3. Thirty-day payment terms | In the total, terms sheet quoted | Section 3 |
| 4. First delivery by 14 January 2027 | In the total, ten weeks from acceptance | Section 5 |
| 5. Named supervisor and a weekly report | In the total, with the report attached | Section 2 |
| 6. Faulty units replaced within 48 hours | In the total, with the clause quoted | Section 3 |
| 7. Public liability, buyer named | In the total, with the limit and the certificate | Section 3 |
| 8. 24-month warranty | As a priced option, $1,380.00, terms quoted | Sections 3 and 4 |
| 9. Installation at both sites | Marked open, $0.00, question with a date | Section 6 |
Seven requirements are answered inside the total, one is answered with a price attached, and one is open. That is the honest distribution for a first draft, and it is more useful to the buyer than a document that appears complete.
The buyer’s reply prices installation at $860.00 for both sites, which takes the total to $23,580.00, and declines the warranty extension, which leaves it at $23,580.00 rather than $24,960.00. The seller sends the same day, one page, with the revised total, the unchanged validity — still 5 November, because the offer moved by one line rather than by the whole document — and the decision requested again, to the named approver, with the date unchanged.
The number that matters is not the total. It is that $860.00 was outstanding on a document whose entire drafting cost was $0.04900: six calls for the proposal and one for the comparison table at the measured Base rate of $0.00700. $860.00 is 17,551 times that cost, and it was a figure that would have become a commitment had the seller estimated it rather than asking. The proposal’s model cost is 0.0002 per cent of the $24,960.00 the buyer would commit if both outstanding items were taken, which is the reason the drafting cost is never the argument for skipping a step and the reason it is worth spending an extra call on.
At the metered ceiling of $0.00955 the same seven calls cost $0.06685. Re-reading the two sections that make a commercial claim — the price table and the terms — on Qwen flash at its measured $0.00875 adds $0.01750, taking the document to $0.06650, which is 35.7 per cent more than the exact-rate figure for a second independent read on the two pages a buyer will read twice.
When the reader is a process
Some proposals are not read by a person who can decide. They are scored against a matrix by a panel, or opened by a procurement function that will not take a call, or assessed against criteria the buyer has published. The structure in Table 1 still applies, and the close has to be rewritten, because the request is now addressed to a process.
In that case section 8 states what the process requires rather than what the seller wants: the date on which the buyer has said it will decide, the format the buyer has asked for, the contact the buyer has named, and what the seller will do on each of those dates. A close addressed to a person inside a process that does not consult people before deciding is a request sent to the wrong address, and it reads as a seller who has not understood the procedure.
There is one thing that does not change. The offer still needs a validity, because a price that holds indefinitely is a price the seller has given away. Where the buyer’s timetable runs past the seller’s validity, the correct move is to say so in the document, with the date and the reason, rather than to quietly extend a date that nobody will check until the contract is drafted.
What structure cannot do
The first limit is that a structure cannot make a weak offer strong. If the price is uncompetitive, the lead time is impossible, or the terms exclude the thing the buyer cares about, eight sections will make that clearer rather than less clear. A perfectly structured proposal for the wrong offer is a faster no, which is worth having and is not a win.
The second is that the proposal cannot answer a requirement it was never given. Where the brief is silent on a material point, the answer is a question — and a question looks like weakness to some buyers, particularly in a scoring process. That cost is real, it is accepted deliberately, and the alternative is a number the seller invents, which is worse in every case that ends in a contract.
The third is that every figure is only as good as its source. The proposal’s discipline is that each number traces to a document the client supplied. Where that document is out of date — an old price list, a warranty schedule that changed, a lead time that moved — the proposal carries the staleness into a commitment, and the trace makes the error easier to find afterwards rather than less likely beforehand.
The fourth is that the decision requested presumes a decision-maker. In an organisation where the reader is assembling a case for somebody else, the close is a request to carry the request, and the most useful thing the seller can do is make that carrying easy: the requirement restated, the total reconciled, the terms summarised, the gaps named. That is a different piece of writing from a pitch, and it is the one that gets a proposal through an internal paper.
The fifth is that negotiation happens after the proposal and is a different document. A structure that tries to win the negotiation inside the proposal — tiered discounts, escalators, early-payment incentives — makes the decision harder, because the buyer now has to model four futures instead of choosing between two offers. The proposal’s job is to make one offer decidable, and the negotiation’s job is the next document.
The sixth is that this piece claims nothing about outcomes. It quotes no close rate, no win rate and no benchmark, because this publication has measured none. What it can say is narrower and still useful: a proposal built this way can be checked, requirement by requirement, before it is sent — and a document that can be checked is a document whose gaps are known to the seller rather than found by the buyer.
The decision
Answer the brief in the brief’s order and numbering, restate the requirement in the buyer’s words at the top, and let each answer sit beside the term that qualifies it. That is the structural decision and it is taken once, before drafting, because retrofitting numbering onto a narrative costs more than writing it correctly.
Then make the close a dated request: what acceptance is, who signs, by what date, what the seller does on each date, and what happens if the answer is no. A close that cannot be refused is pressure; a close that can be refused, with the date stated and the file left open, is a request — and it is the only version of the close that a buyer can act on without being managed into it.
And carry nothing that does not trace. Where the source is silent, the marked gap and the question, never a plausible number — because a number the buyer reads as an offer becomes one the moment it is sent, and $860.00 of installation estimated rather than asked would have been a commitment made by a seller who was trying to be helpful.
Where this goes
- The pillar: x402.review/2026/10/31/commercial-operations/ — one record, two readers, and the approvals between them
- The role that drafts it: studio.x402.press/roles/sales — the proposal, the comparison table and the figures it may not invent
- The table the buyer will build: the three-quote comparison template — the columns that decide which offer wins
- The basis a proposal has to survive: comparing quotes that are not like for like — normalisation, and what it does to a quoted price
- The service it buys: x402.press/services/deepseek/ — the canonical price, seven chains
- The review: x402.review/2026/10/31/deepseek-review/ — the measured handshake and what it does not promise
Method note
The price was read from the live 402 challenge on Base on 31 October 2026, a Saturday and therefore outside DeepSeek’s weekday peak windows: 7000 atomic units under exact, which is $0.00700, and 9550 under upto at the max_tokens tested. Qwen flash’s $0.00875 was read from its own challenge the same day. Call counts follow the sales role’s specification: six calls for a proposal, one for a comparison table.
Meridian Care Group, the brief, the price list, the terms sheet, the dates and every amount in Tables 2 and 3 are constructed for this piece; no real buyer, offer or contract is described. The structure in Table 1 is the discipline the sales role is specified to follow, not a tested template with a measured effect. No close rate, win rate or timing benchmark is quoted, because none has been measured.
x402.review is written and edited by N.P. Vincent. The publication is connected to x402.press, which sells the services reviewed here; that interest is disclosed on every piece that touches a service, and it is stated here rather than left to inference.
Filed under Jobs · Tags: sales, proposals, structure, decision
