How to Write a Proposal That Survives an Approval Chain
Why this matters
A residential proposal has one reader who was standing next to you when you found the problem. An institutional proposal has four or five readers, each further from the equipment than the last, each with less time than the last, and each with a different question. The document that dies is almost never the one with the wrong price. It is the one that answered the first reader's question thoroughly and left the last reader unable to answer theirs in ninety seconds. So write it backwards: start from the reader who is furthest away, and work down toward the mechanic who already agrees with you.
Who reads it, in the order they matter to the outcome
The chain below runs from the last reader to the first. That is the reverse of the order the document travels, and it is deliberately how you should build it.
The board member or executive committee member. Sees an agenda summary, often a paragraph. Their question: what happens if we do not do this, and was it bought properly? They are not evaluating your scope. They are deciding whether to expose themselves to a decision.
The business manager or controller. Their question: which fund, which account, and does it fit the threshold band? They will not chase a scope to work out whether it is capital or operating. If they cannot tell in one read, they send it back rather than guess. See related: Capital Money and Operating Money Are Not the Same Money.
Purchasing. Their question: is this procurable as written? Are you a registered vendor, is your certificate of insurance current with the right additional-insured wording, and is the scope written so a competing quote can be compared to it line for line.
The facilities director. Their question: what does this do to my variance and my complaint count, and what happens if I defer it a year? See related: The Facilities Director and What They Are Actually Judged On.
The maintenance supervisor. Their question: is the diagnosis right and is the scope complete? This is the reader shops write the whole document for, and the one whose agreement you most likely already have.
The attention arithmetic that should shape the page order
Count the real reading time a fourteen-page proposal gets. Roughly: supervisor 20 minutes, director 10 minutes, purchasing 4 minutes, business manager 5 minutes, board member 1.5 minutes. That is 40.5 minutes of human attention across the five, and the reader with the final vote spends 1.5 of it, which is under 4 percent of the total.
Everything that reader needs therefore lives on page one, in prose, with no attachment reference. Not because they are unserious, but because they are deciding forty items that night and yours is one of them.
Build it from the bottom, write it from the top
Gather from the bottom. The supervisor gives you the fault, the history, the failure dates, and the constraints. Ask them for the failure dates specifically; a documented failure history is what converts an opinion into a risk statement four readers later.
Write from the top. Draft the one-paragraph summary first, before the scope. If you cannot write a paragraph that states the risk of inaction, the recommendation, and the fund, you do not yet understand the job well enough to price it, and everything downstream will inherit that fog.
What page one must contain
- The condition, in one or two sentences, with dates. Not adjectives. "Unit B failed on a stated date, has not been returned to service, and Unit A has carried the full load since."
- The consequence of doing nothing, stated as an event, not a feeling. What breaks, who it affects, how many of them, and when it becomes likely.
- The recommendation, one option named as recommended. Give two or three options, and say plainly which one you recommend and why. A proposal with three equal options outsources the decision to someone with less information than you, and they will pick the cheapest or defer.
- The fund classification you believe applies. State whether you have written this as operating or capital work.
- The threshold band you believe it falls into, and what that band requires. Buyers respond well to a vendor who already knows their rules. See related: How an Institution Decides to Spend Money.
- The schedule constraint. Outage window required, equipment lead time in weeks, and the date after which this cannot be completed inside the current fiscal year.
Everything else is an attachment: photos, readings, line-item pricing, the certificate of insurance, the vendor registration reference, references from comparable institutions.
What kills proposals that are otherwise good
A price with no basis of comparison. In the informal-quote band your document will be set beside two others. If yours is written as a lump and theirs are itemized, yours is the one that cannot be evaluated, and the safest thing for a buyer to do with an unevaluable quote is set it aside.
Options that are not really options. Three tiers of the same work with escalating trim is a residential move and it reads as one in an institution. Real options are structurally different answers: repair, replace this asset, replace the plant, or defer with a stated monitoring plan.
No deferral option. Institutions defer constantly. If you have not written what deferral looks like, including what you would monitor and what would trigger a return visit, someone else will invent a deferral plan that is worse than yours.
Bundling capital and operating scope on one document. It cannot be paid without being split, and the split costs weeks.
Assumptions buried at the back. Anything that changes the price if it turns out to be wrong belongs on page one under a heading that says so. Concealed conditions, access assumptions, the presumption that existing electrical service is adequate.
A proposal traced through the chain
A university residence hall with two gas-fired domestic hot water heaters serving about 400 residents, arranged lead and lag with each sized to carry the full peak load. One failed in late spring and has not been repaired. The survivor has carried the whole load since.
What the supervisor gave us. Failure date. The survivor's runtime, which has roughly doubled from carrying about half the annual load hours to carrying all of them, with burner cycles up in proportion. Two prior repairs on the failed unit in the preceding three years.
Reading the exposure at both ends of the year, because that is what makes the risk statement honest. Over the summer, with the hall near empty, the survivor is lightly loaded and the only real loss is redundancy. On a mid-semester winter morning it runs continuously with no margin at all, and a failure then is not reduced service, it is no hot water in an occupied residence hall, which is a habitability event with relocation attached. If the exposure were worst in summer, the story would be about capacity rather than redundancy and the recommendation would change.
Page one, written first. Condition with dates. Consequence: one component failure away from a hot water outage affecting roughly 400 residents, worst case in mid-semester winter. Three options: repair the failed unit, replace the failed unit like for like, or replace the full plant. Recommendation: replace the failed unit, because it has needed two repairs in three years and a third restores nothing but the same exposure. Fund: replacing one unit is likely capital if it clears their capitalization threshold, so it is written that way and flagged for the business manager to confirm. Band: stated. Schedule: the outage window is the winter break, and equipment lead time in weeks means a purchase order is needed by a named date to make that window.
How each reader used it. The supervisor confirmed the diagnosis and added that the flue on the failed unit was oversized for a single appliance once the pair became a single, which nobody had noticed. The director had the risk sentence and the resident count, which is what they needed for their own escalation. Purchasing found the scope comparable enough to solicit against. The business manager confirmed capital and coded it. The board member read one paragraph containing a number of residents, a season, and a named recommendation, and voted.
Where it nearly died. The first draft opened with combustion analysis readings and put the resident count on page six. That version was returned by purchasing with a request to clarify the fund, which would have cost a board cycle, roughly a month, and pushed the work past the winter break window and into a summer that was two full semesters away.
The full-plant replacement option matters even though it was not recommended. Naming it, with a rough magnitude in relative terms and a note that it belongs in the capital plan rather than in this request, is what puts the plant on next year's budget development agenda. See related: The Budget Cycle and Why Timing Can Beat Price.
The safety the recommendation itself creates
Two things in this job put a hazard on residents rather than on your technician, and both belong in the scope document, not just in the technician's head.
Do not raise storage temperature to compensate for lost capacity while a single unit carries the load. Higher delivered water temperature at a fixture used by residents is a scald exposure, and the control is thermostatic mixing at the point of use with delivered temperature verified at fixtures before the system is returned to occupied service, not a lower storage setpoint. Storage temperature and delivered temperature are two different numbers and they are managed by two different devices.
Do not lower storage temperature either, as an informal fix. Storage held in the range that favors bacterial growth in a large building water system is exactly what a water management program under ASHRAE Standard 188 exists to prevent, and a residence hall is a building where that program should already exist. Coordinate any temperature change with whoever owns it.
On the mechanical work itself: the storage vessel is isolated at the fuel supply and the electrical supply, allowed to cool, and vented with the vent verified open before it is drained or opened, because draining a hot vessel is a scald exposure to anyone near the drain path, and the stored energy isolation runs under 29 CFR 1910.147. The electrical disconnect work is not covered by that standard, which excludes exposure to electrical hazards from work on conductors and equipment in electric utilization installations at (a)(1)(ii)(C); it runs under 29 CFR 1910.333(b)(2), with conductors proved dead using an instrument checked on a known live source before and after, per NFPA 70E-2021, 120.5.
How to verify the proposal will survive before you send it
Cover everything except page one and hand it to someone in your office who has never seen the job. Ask them three questions: what happens if this is not done, what are we recommending, and which pot of money pays for it.
If they can answer all three in under two minutes, the board member can too. If they cannot answer the first one, no amount of technical quality further in the document will save it, because the first reader with authority to defer will defer.
Then check one mechanical thing: that the date after which the work cannot be completed this fiscal year appears on page one, in numerals, as a date. That single line is what moves a proposal from a queue into an agenda.
References
- 29 CFR 1910.147, control of hazardous energy for stored energy in a fired storage vessel, with the electrical utilization exclusion at (a)(1)(ii)(C); 29 CFR 1910.333(b)(2) and NFPA 70E-2021, 120.5 for the electrical isolation and verification
- ASHRAE Standard 188, legionellosis risk management for building water systems, which owns storage temperature decisions in a large occupied building
- See related: The Facilities Director and What They Are Actually Judged On; Capital Money and Operating Money Are Not the Same Money; Structuring a Long Proposal So It Gets Read