Structural and civil engineering design, drainage and surveys for developers, contractors, schools and industrial clients across the West Midlands and the Welsh Marches. 48 staff; five people write proposals.
Verdict: The log holds 52 proposals; they average 17 pages and, on the types in the scoring analysis, score 3.1 out of 10. The points met least often are Within the page limit for its type (3%); Opens with the client's situation (16%); Offers options (18%); 14 of 52 proposals failed the gate (answers everything the client asked). By score band, won within the window ÷ sent: Low (0 to under 4): 27% of 22 (give or take 19 points); Middle (4 to under 7): 29% of 7 (give or take 34 points); High (7 to 10): none. Checklist points whose win-rate difference is beyond chance: 2 of 7 (association, not proof). Other things go with winning too: Repeat client: too few with (4), 28% without (25); Competitive: 26% with (19), 30% without (10); Pre-proposal meeting: 45% with (11), 17% without (18). Beyond chance: 0 of 2. The new template has 11 sections, 9.5 pages counted (shortest limit 10.0); a seven-part executive summary of 390 words; approved by Rhys M on 14 Oct 2026. Primary measure: comparable proposals take 16.5 hours to write (give or take 1.7); the owner's target of 10 hours (assumed) can be read at the re-measure if the change exceeds 2.5 hours. On comparable proposals the business wins 28% within 90 days (give or take 17 points, from 29 proposals); other quotations counted the same way win 38%. No decision: 38% at 90 days (of 29); 29% at 120 days (of 28); 26% at 180 days (of 23); a decision takes 58 days on average, with 8 still open. At this volume the win rate can only show direction, not size: the smallest change it could show (23 points, at 80% power) is wider than the gap to the owner's target. Measure again on 16 Oct 2027, the same way (same types, source, window and scorers, scoring blind); a change in the win rate under 31 points could be chance. The early read on 14 Jan 2027 re-scores new proposals mixed with old, dates hidden, and compares writing hours. For scale only (illustrations, not forecasts): the hours target would be worth £8,249 a year and the win-rate target £51,625 in gross profit, against £6,944 of fee and team time. Not yet measured: fill in the Value Case after 14 Jan 2027.
Red: the three points met least often. Grey: point 10, the pass/fail gate; 14 of 52 proposals (all types) failed it. The proposals average 17.3 pages; only 5 of 52 fit the new page limit for their type. Most open with Callowdene’s history, not the client’s project, and few offer a choice. Tenders and surveys are left out of this analysis: tenders follow the buyer’s format and surveys are short fee letters. N/A is allowed where a point can’t apply.
| Band (scores) | Proposals | Sent 90+ days ago | Won within 90 days | Win rate | Give or take | Repeat clients | Pre-proposal meeting |
|---|---|---|---|---|---|---|---|
| Low (0 to under 4) | 26 | 22 | 6 | 27% | ±19 | 18% | 45% |
| Middle (4 to under 7) | 12 | 7 | 2 | 29% | ±34 | 0% | 14% |
| High (7 to 10) | 0 | 0 | 0 | – | – | – | – |
The two bands win at much the same rate, with a give or take wider than any gap, and no proposal scored 7 or more. Of the 7 checklist points with enough proposals either side, 2 showed a difference beyond chance (relevant case study; price clear with inclusions and exclusions); with this many comparisons one or two can be chance alone. The checklist is adopted because it is what a buyer can check in a proposal, not because this log proves it works.
| Factor | Win rate with (sent) | Win rate without (sent) | Give or take | Average score with / without |
|---|---|---|---|---|
| Repeat client | too few (4) | 28% (25) | – | 1.9 / 3.1 |
| Competitive | 26% (19) | 30% (10) | ±35 | 2.8 / 3.2 |
| Pre-proposal meeting | 45% (11) | 17% (18) | ±35 | 2.5 / 3.2 |
Proposals after a pre-proposal meeting were won far more often, though the give or take is wide; repeat clients were too few to show. These sit alongside the checklist: a better proposal can’t replace the meeting, and a change in the mix of repeat or competitive work between now and the re-measure would move the win rate whatever the template does.
| Type | Logged | Average value | Pages now | New limit | Writing hours | Score now | Win rate | Measure |
|---|---|---|---|---|---|---|---|---|
| Structural design | 15 | £19,400 | 18 | 10 | 16.5 | 3.7 | 18% | Measured |
| Civil and drainage design | 11 | £17,045 | 16 | 10 | 14.5 | 2.5 | 22% | Measured |
| Structural and civils together | 12 | £55,313 | 22 | 12 | 18.2 | 3.0 | 44% | Measured |
| Surveys and reports | 9 | £3,917 | 6 | 4 | 6.7 | 3.1 | 14% | Not measured |
| Public tender | 5 | £78,950 | 27 | 25 | 34.7 | 5.1 | too few | Not measured |
Public tenders follow the buyer’s own format and scoring under the Procurement Act 2023, so they are left out of the measure (whether to bid at all is the Tender Bid Decision). Surveys and reports get a short version of the template but are not measured. Win rates show only with five or more proposals sent at least 90 days ago.
| Section and purpose | Page limit | Checklist point it serves | Who writes it |
|---|---|---|---|
| 1. Cover and headline Say in one line what changes for the client | 0.5 | – | Rhys M |
| 2. Executive summary Let a busy decision-maker decide from one page | 1 | Executive summary of a page or less | Rhys M |
| 3. Your situation and what you need Show we have understood their problem before we talk about us | 1 | Opens with the client's situation | Anita K |
| 4. What you will get State the outcome in terms the client can measure | 0.5 | States a measurable outcome | Anita K |
| 5. Options and our recommendation Give two or three ways to proceed and say which we recommend | 1.5 | Offers options | Rhys M |
| 6. How we will do it Approach and programme against everything the client asked for | 2 | Answers everything the client asked | Tom B |
| 7. Risks and how we handle them Name the risks the client worries about and what we do about each | 0.5 | Risks named with how they are handled | Anita K |
| 8. Why us: a similar job One relevant case study and the people who will do the work | 1 | Relevant case study | Gemma L |
| 9. Investment Fees ex VAT, what is included and excluded, payment terms, validity | 1 | Price clear with inclusions and exclusions | Paul D |
| 10. Next step and decision date The one thing we ask them to do, and the date | 0.5 | Next step and decision date | Rhys M |
| 11. Appendices Supporting material the reader can skip | 0 (not counted) | – | Gemma L |
Every checklist point except the page limit has a section that delivers it. For an engineering consultancy the risks section covers how the firm meets the designer competence duties in Part 2A of the Building Regulations where it applies, and the investment section sets out stage payments in line with the Construction Act and the professional indemnity cover. Not legal advice.
| Before (104 words) | After: seven parts (196 words) |
|---|---|
| Callowdene Consulting Engineers was established in 1998 and is a multi-disciplinary consultancy offering structural, civil and geotechnical engineering services. We have a team of 48 staff including chartered engineers and technicians, and we are committed to delivering high-quality, innovative and cost-effective solutions for our clients. We have extensive experience in the industrial and commercial sectors. Our services for this project would include structural design of the extension frame, foundation design, ground floor slab design, drainage design and attendance at design team meetings. Our fee proposal is attached at Appendix C. We trust this proposal meets your requirements and look forward to hearing from you. | The client's situation. Your new racking arrives in March, and the warehouse extension has to be built and its slab cured before it does. Picking in the existing building cannot stop while it happens. What they want. You want a slab that takes the racking loads first time, drawings your contractor can price from by 4 December, and no surprises at Building Control. Our recommendation. We recommend Option B: we design the frame, foundations, slab and drainage, and agree the slab joint layout with your racking supplier before the drawings are issued, so the racking legs never land on a joint. Why us. In 2025 we designed a 4,200 m² slab extension for a chilled-food distributor in the same way; it passed Building Control first time and the racking went in eleven days after handover. Their operations manager has agreed to take a call. The outcome. Tender drawings issued by 4 December, Building Control approval before the contractor starts, and a slab designed for your racking supplier's leg loads. Investment. £18,750 ex VAT (VAT £3,750), in three stage payments, including two site visits and Building Control liaison. Next step. A 30-minute call with your racking supplier this week; we need your instruction by 6 November to hold the 4 December date. |
The “before” opens with the firm’s history, lists services and ends by waiting to hear. The “after” opens with the client’s deadline, recommends one option, gives one checkable result from a similar job with a contact, states the fee ex VAT with what it includes, and asks for one action by a date. It is longer in words, but it is the page a decision-maker reads; the firm’s history moves to an appendix.
Primary: writing hours. From the fee-bid timesheet code, 38 structural and civils proposals took 16.5 hours each on average (give or take 1.7). A ten-page template should cut that; by the re-measure about 30 new proposals will have hours, so a change larger than 2.5 hours will show. The owner’s target of 10 hours is an assumption.
| Window | Sent at least that long ago | Won within it | Win rate | Give or take | No decision | Still open now |
|---|---|---|---|---|---|---|
| 90 days | 29 | 8 | 28% | ±17 | 38% | 8 |
| 120 days | 28 | 9 | 32% | ±18 | 29% | 8 |
| 180 days | 23 | 8 | 35% | ±20 | 26% | 6 |
Then the win rate, no-decision share and days to a decision. Proposals never answered count as not won. Other quotations, counted the same way, win 38%; the re-measure nets off their change. A decision takes 58 days on average across every decided proposal (give or take 10), with 8 still open and reported separately. At about thirty comparable proposals a year the win rate can only show direction, not size: the smallest change it could show (23 points, at 80% power) is already wider than the gap to the owner’s 40% target, and the change by the re-measure has a give or take of ±31 points. The early read on 14 January 2027 re-scores new proposals mixed with old ones, dates and outcomes hidden, by the same two people, and compares writing hours. The re-measure uses the same types, source, 90-day window and scorers.
For scale only: the hours target would be worth £8,249 a year at £42 an hour; a 10-point rise in the win rate £41,587 in gross profit. Illustrations, not forecasts. Fee and team time, built up from 108 hours of scoring, the workshop, the template and blind re-scoring: £6,944.
Proposals come mainly from Proposal folder (4% from won-only files) and outcomes from Quotation log (won and lost, with dates), checked against email, over 15 months; the log holds 95% of the proposals sent. Each proposal was scored by Anita K and the consultant together; the score beside the outcome is an association, not proof, and repeat clients, competition and pre-proposal meetings move win rates too. Writing hours rest on 38 proposals with hours recorded, give or take 1.7 hours. The win rate counts proposals of the measured types sent at least 90 days before the date (won within 90 days ÷ sent, so unanswered ones count as not won); it rests on 29 proposals, give or take 17 points, and the change by the re-measure has a give or take of 31 points including the background's own variance. The targets (10 hours, 40%, 40 days) are the owner's assumptions, not forecasts; the won value (£36,000) and margin (38%) come from the records and count one year only. The re-measure nets off the change in other quotations counted the same way, and shows whether the measures moved, not proof that the template moved them.
Writing hours are the measure most likely to move and be seen within the year. The win rate will show direction; the no-decision share and days to a decision will show whether the next step and decision date are working. No client names are held in the log: references only.
Fictional, continuing the example. After the template was approved, Callowdene chose the guided route and set up the agent in an AI assistant the business already uses, connected read-only to the CRM, the enquiries inbox and the timesheets' fee-bid code, and able to save drafts to the proposals folder; its IT confirmed this before it committed. Callowdene updated its privacy notice first, since enquiries carry its clients' contact details. For each enquiry marked as needing a proposal it drafts one to the approved template, 3 working days before it is due, for the writer to finish; it runs the gate and the review checklist on the finished version; and it drafts a follow-up for the writer when a proposal has had no reply for 14 days. Nothing left the business, and no fee or term changed, without one of its own people approving it: writers finished and sent every proposal from their own email, after a director approved it.
| Date | What the agent prepared or found | Who decided, and what happened |
|---|---|---|
| 19 Oct Decision | First draft to the template for CD-2610-53 (structural design: a mezzanine floor in a distribution warehouse), due 23 October | Anita K (writer); review checklist by Paul D; approved by Rhys M: Sent on 22 October after the writer's edits; gate passed; 9 writing hours booked to the fee-bid code |
| 27 Oct Decision | Gate not passed on CD-2610-55 (civil and drainage design), finished and ready to send | Paul D (writer); approved by Rhys M: Answer added; sent on 28 October |
| 5 Nov Decision | Follow-up drafted for the writer: CD-2610-53 had no reply 14 days after sending (Inputs B44) | Anita K: Sent from her own email on 5 November; instruction received on 24 November |
| 9 Nov Decision | A case study for CD-2611-58 would name the client's contact: no written permission on file | Gemma L: Written permission received on 12 November and recorded; the contact is named in proposals from then on |
| 30 Nov Note | Weekly note: proposals drafted to the template since 16 October, with writing hours | No decision needed |
| 4 Jan Decision | Blind early read prepared for the scorers | Anita K and the consultant: Scoring booked for 13 January. The scorers, not the agent, enter the early read on the Value Case |
The check-in on 18 November, 30 days after its first draft, read the log with Rhys M; after that the agent was Callowdene's, and Geoff kept no copy and no access. Writing hours on the first six proposals averaged 11.5, against 16.5 at the baseline, but six proposals are too few to judge: the give or take on the change at the re-measure is 2.5 hours on a full year's proposals. The early read on 14 January is scored blind by the same people; the re-measure on 16 October 2027 is the one to judge. No £ is claimed for the months after, and they add nothing to the figures above.