ingridleiria Proposal Writer
Most proposals are lost before the price page. They are lost on the first page, where the client reads a description of the firm instead of a description of themselves and concludes the document was sent to everyone. From there the reader skims for the number, which has nothing attached to it but generic capability, so it looks expensive. The deal is rarely rejected; it goes quiet, which costs more, because a quiet deal stays in the forecast and keeps consuming attention.
The second failure is slower. A proposal that wins but describes the work loosely becomes the reference document for an engagement nobody has scoped, and every ambiguity gets resolved in the client's favour for six months. One vague deliverable, "a report on the current state", costs more in unbilled rework than writing the whole proposal properly would have cost.
When to use this, and when not to
Use it when a named client has to decide whether to buy work: a first proposal after discovery, a follow-on phase, a renewal, a retainer conversion, an RFP or tender response, or a rescue of one that came back with objections.
Do not use it before discovery has produced the client's own statement of the problem and of what success looks like. A proposal written from your diagnosis alone is a guess with a price on it, and the right output is a list of questions and a call.
Adjacent cases. Presenting an approach for reaction, before anyone is asked to buy, is discovery-to-proposal-deck when the raw material is one client's discovery notes, or ideation-deck when the idea has no named buyer yet. Rebuilding a standard deck around one account, including co-branding, is tailored-client-deck; a proposal and a deck often ship together and are different documents. The price behind the number is pricing-and-resourcing-model, the document that governs the work once they say yes is sow-and-scope, and papering an associate onto what you sold is contractor-msa-and-task-order. Agreements that are not client scope, including referral papers, are business-agreements-drafting. Asking an investor for money is pitch-deck.
What you need before starting
The client's problem in their own words, with their numbers. Quotes or close paraphrase from discovery, not your restatement. Missing: draft the discovery questions instead, because without this the document can only be a template and the client will recognise it.
Their definition of success and how they would measure it. Missing: propose two candidate definitions in the covering email and ask which is closer. Never invent an outcome metric, because you will be held to it.
Who decides, who influences, who can block. The economic buyer, the champion who carries it internally, and the function that says no late: procurement, legal, security, or finance. Missing: ask the champion in those words, and default to writing for them to forward.
The offer: approach, deliverables, team, duration, price and structure. From pricing-and-resourcing-model. Missing a price: send the approach for reaction and follow with the price, and never state a range you have not built bottom up.
Budget signals and procurement rules. The threshold above which a competitive process starts, purchase order requirements, payment terms they will not vary. Missing: ask what the approval path looks like above a given figure. That one question prevents the most common late collapse.
What competes with you, and the cases you can prove. Competition includes doing nothing and doing it internally. Missing: assume internal delivery is the live alternative and answer it inside the approach. Use only cases you can describe with a figure and a date.
The method
Fix the decision and the reader. One line in your notes: who signs, what they sign, by when. Everything is then judged on whether it helps that person say yes or helps the champion defend it. Where the two need different things, write for the champion and put price and terms where the approver finds them without reading the rest.
Write the situation section first, from their language only. No firm content of any kind in it. The test: delete the client's name and show the section to a colleague. If they can guess only the industry, it is not specific enough; if they can guess the company, it is right.
Convert the situation into two to four outcomes. An outcome is a state of the world after the work, with a measure. "Opportunities qualified at stage two above eighty percent by week twelve" is an outcome; "improved sales effectiveness" is a slogan. Rule for how many: one per problem the client actually stated. Extra outcomes are your ideas and belong in an appendix or nowhere.
Describe the approach by what each phase produces. Each phase gets a plain name, a duration, what the client's team does, and the artefact that exists at the end. Rule: enough phases that the client can see a point where they could evaluate and continue, and no more, which is almost always three or four. One monolithic phase makes the fee feel like a single irreversible bet.
Write deliverables as though a dispute has already started. Name, one-line description, format, and one short relevance paragraph per named person on the team rather than a biography. Then exclusions, wherever a reader might assume more: implementation after recommendations, training beyond a stated number of sessions, data cleansing, support after the end date. Rule: if you would be annoyed to be asked for it in month three, exclude it now, where it is a scoping detail rather than a refusal.
Build the timeline against their calendar and name the dependencies. Milestones as dates or weeks from kickoff, and beside each the client-side input that holds it: a named person's time, data by a date, a decision turnaround. This is what makes the plan credible, and it sets up the change conversation before you need it.
State the investment plainly, under its own heading. Fee, structure, payment schedule, included, not included. Rule on options: at most two, and only where they are different levels of commitment rather than a discount ladder. Never split the commercial picture across three places.
Put credibility late and attach a figure to every claim. Two or three reasons tied to this problem, one case each with a number and a date. Rule on placement: nothing about the firm appears before the investment section. Credentials read before the price are read as justification for the price and get discounted; credentials read after it are checked, which is what you want, because the reader has already decided whether the work is worth buying and is now asking only whether you can do it. The single exception is a tender pack that demands firm background in section one, where the buyer controls the order. Rule on which cases: choose them to match on whatever made this problem hard, not on industry label. A same-sector case with a different difficulty proves nothing and invites the comparison you would lose. No superlatives: "world-class" asks the reader to take on faith the one thing they were about to check.
Close with one next step, dated and owned. "Countersign by the twentieth and we start the week of the twenty-seventh" is a next step; "let us know your thoughts" is an invitation to defer.
Run the two-reader test. Read once as the champion, who must forward and defend it without you present, and once as the approver, who has ninety seconds and wants the number, the term, and the risk. Anything neither needs goes to the appendix or out.
For a formal RFP or tender these rules are overridden where they conflict with the pack. Build a compliance matrix first, follow the requested structure and numbering exactly even when it is poor, answer each question in its first sentence, and move the situation content into the executive summary. Non-compliance on format disqualifies more good responses than weak content does, so check page limits, file formats, naming conventions, submission channel, and the deadline including its time zone.
Worked example
Situation. Arbor Analytics, a twenty-two person analytics firm, ran two discovery calls with Northbank Logistics, a regional freight company of four hundred staff. All figures in this example are US dollars. Northbank's operations director said depot managers made next-day capacity decisions from a spreadsheet one analyst rebuilt every evening, that the analyst had resigned with six weeks of notice, and that they had lost roughly 180,000 dollars the previous year on empty return legs. Their finance director had said anything above 75,000 dollars goes to a three-supplier comparison.
Task. A proposal within six working days that the operations director could forward without a meeting. Good meant a signature or a clear no before the analyst left.
Action. The situation page was written first from call notes only: the resigning analyst by role, the nightly rebuild, the 180,000 dollars attributed to them, and their own phrase, "we plan the morning after the day we needed to". Nothing about Arbor appeared on it.
The first version of the offer was a twelve-week rebuild of the planning process at 96,000 dollars. It was abandoned once the procurement threshold surfaced. At 96,000 dollars the proposal triggers a three-supplier comparison that would run longer than the six weeks of notice remaining, so Northbank would lose the knowledge before any supplier started. The threshold was not an obstacle to work around; it was information about what they could buy.
The rewrite split the work. Phase one at 48,000 dollars over eight weeks, ending with the nightly model running as a scheduled job the depot managers could trigger themselves, a written handover of the analyst's logic, and one training session. Phase two, the wider planning redesign, was named and sized as a range to be scoped separately. Outcomes were cut from five to three, because Northbank had stated three problems and the other two were Arbor's ideas. The timeline named the two dependencies that could break it: database access by end of week one, and four hours of the analyst's time in weeks two and three.
Result. Sent on the Wednesday and approved without a call, because it sat under the threshold and payment was two milestone instalments. Kickoff was eleven days later with two weeks of notice remaining. Phase two was signed four months on at 71,000 dollars, and that conversation started from the phase one deliverable list rather than from nothing. What made the difference was treating the procurement threshold as a design constraint on the offer rather than an administrative detail for the terms section.
A second scenario, where it goes differently
The same firm answered a formal tender from a public transport authority, and almost none of the above applied. The pack specified eleven sections in a fixed order, a forty-page limit, no appendices, and a published rubric weighted sixty percent method and forty percent price.
The situation content compressed into a nine-line executive summary, because section one was defined as "organisational background". Two case studies that would have carried a commercial proposal were cut, because the rubric gave them no points and the page limit was binding. The response scored well on method and lost on price to a larger bidder. What changed was not the quality standard but who controls the structure, and when the buyer controls it, compliance outranks every instinct about how a proposal should read.
Output
COVER client name as they spell it, engagement name, date, prepared by, valid until
1 YOUR SITUATION one page, their words, their numbers, no firm content
2 OUTCOMES 2 to 4, each with a measure, a target, a date
3 APPROACH phases, each with duration, client role, what it produces
4 DELIVERABLES table below, plus an explicit exclusions list
5 TEAM one relevance paragraph per named person
6 TIMELINE milestones, each with the client input it depends on
7 INVESTMENT fee, structure, payment schedule, included, not included
8 WHY US 2 or 3 reasons, one case each, every case with a figure
9 TERMS validity, payment terms, IP, confidentiality, cancellation
10 NEXT STEP one action, one date, one owner
APPENDIX full bios, case detail, method notes, references
| # | Deliverable | Description | Format | Phase or date | | 1 | | | | | | 2 | | | | |
Exclusions follow the table as plain sentences: "This engagement does not include ..."
Failure modes
The firm's story on page one. Recognise it because the first noun in the document is the firm's name. The reader classifies it as marketing in ten seconds and never fully reclassifies it. Rewrite page one from discovery notes only.
Activities dressed as outcomes, and deliverables too vague to check. Recognise the first when a deliverable and an outcome say nearly the same thing, for example "workshops delivered" in both lists; the outcome is what is true afterwards that was not true before. Recognise the second by asking whether someone outside the engagement could tell "a report" or "support during rollout" had been delivered. Fix both by adding a format, a length, and a boundary.
Three prices as a hedge. Recognise it when options differ only in fee and volume rather than in what the client commits to. It hands your decision to the client and often trips procurement rules one price would not.
Claims without numbers. "Deep expertise", "proven methodology". They would survive being pasted into a competitor's proposal unchanged. Replace each with a case, a figure, and a date. The related error, a previous client's name left in the file, is fatal and entirely preventable: search for it before sending.
A timeline with no client dependencies. Every milestone depends only on your team, so when it slips on their side the slip becomes your unpaid extension. Ending with "we look forward to hearing from you" is the same error at the other end of the document: recognise it because the last section contains no date.
Edge cases
No discovery happened and there is no time for any. Say so. Open with your understanding labelled as an assumption, list the three questions whose answers would change the approach, and price a short paid diagnostic as phase one. A proposal that prices its own uncertainty is more credible than one that guesses confidently. Where the missing discovery is itself a signal that the request exists only to satisfy a procurement rule while they buy elsewhere, alongside a very short deadline and requirements copied from a competitor's feature list, decide whether to bid at all; if you bid, write it in half the time and build no custom material.
An incumbent holds the relationship. Do not attack them. Build the proposal around the thing the incumbent structurally cannot do, and price a first phase small enough to run alongside rather than instead of. The decision you are asking for is "try this", not "replace them". Where the stated budget is far below the work required, the same instinct applies: do not shrink the whole scope to fit, propose the largest coherent piece that stands alone, and name what it excludes and what that would cost.
They ask for slides instead. Produce both: the deck for the conversation, the document as the record. Use tailored-client-deck for the deck and keep deliverables, price, and terms identical in both, because the discrepancy is the first thing a careful approver finds.
Quality bar
- Section one could only have been written for this client, with their words and at least one of their figures.
- The price and the next step are each findable in under a minute by someone who has not read the document.
- Every deliverable has a format and is specific enough to settle a dispute about whether it was delivered.
- Exclusions are written explicitly, not implied by omission.
- Every outcome has a measure, a target, and a date.
- Every claim about the firm carries a figure or a named case.
- Every milestone that depends on a client input says so, with the input named.
- No previous client's name survives anywhere in the file.
Adapting this to your context
Written for small and mid-size professional services firms selling engagements of tens to low hundreds of thousands to a named buyer inside a company. The structure survives most settings; several of the settings do not.
- The ten-section order, with credibility late. That assumes you control the document. A tender, a framework call-off or a grant application does not, and there the compliance matrix comes first and their numbering wins.
- The procurement threshold as a design constraint. The buyer's constraint may instead be a board approval date, a capital and operating expenditure split, or a fiscal year end. Design the phase boundary against whatever the answer is.
- At most two options. Some public and regulated buyers require a fixed number of priced variants. Give them what the pack demands, and make the variants differ in scope rather than in discount.
- Phases of a few weeks each. Where a single procurement, clinical or academic cycle is the smallest unit that can be evaluated, make the phase that long and say what interim evidence appears inside it.
- What not to change. Section one is written from the client's own words and numbers with no firm content in it, and the document ends in one dated, owned next step.
Related skills
discovery-to-proposal-deck produces the approach deck that usually precedes this document and supplies the situation and objectives it is built from, and ideation-deck covers the earlier stage where an idea has no named buyer. pricing-and-resourcing-model builds the fee this document states, and tailored-client-deck builds the presentation that accompanies it. sow-and-scope turns the accepted proposal into the document that governs delivery, and any discrepancy between the two becomes the client's argument; contractor-msa-and-task-order papers the associates who staff it, and business-agreements-drafting covers agreements that are not client scope. outreach-email writes the covering message, principal-simulator and sales-call-analysis pressure-test it against the objections the buyer will raise, and pitch-deck is the investor equivalent.