More proposals, fewer wins

The numbers, and which firms they describe

Win rates get quoted without their population, so here are the three that matter with theirs. Deltek’s 46th Clarity study put the median win rate at 50% across all architecture and engineering firms in its sample, and recorded proposal volume down 38% and the value of awarded work up 52% (Deltek, 13 May 2025). In other words, 2025 was a year of fewer, larger wins. Deltek’s FY2025 data gives architecture firms specifically a 45% win rate, and that is the figure our firm health scorecard scores you against. Then the 47th study, covering FY2025 across 896 firms, found that firms had submitted 32% more proposals while win rates declined; Deltek did not publish the new rate (Deltek Clarity podcast recap, 9 July 2026).

Read the sequence as a cost curve. Proposal effort is unbilled senior time. When volume rises 32% and the win rate falls, the hours spent per fee won rise faster than either number on its own.

Why this hurts small firms most

Backlog is the reason. Firms billing under $250,000 saw their backlog fall from 4.9 months to 3.1 months in Q2 2026, the sharpest drop of any size group, against 6.3 months for all firms (AIA, 22 July 2026). With three months of work in hand, the owner writes proposals in the evening and the lost ones are never costed.

The arithmetic is simple and worth doing once. At a 45% win rate, for every proposal you win you write a little over one that loses. If a residential proposal takes a principal a day to prepare, each win carries more than two days of unbilled time, before the interview and the follow-up. That is arithmetic from the published rate, not a survey finding, and the number for your own firm comes from your own log.

It also explains why busy and profitable come apart. A firm that answers a thin pipeline with more proposals accepts more marginal projects, and marginal projects are where write-downs live. Our article on where the margin leaks covers the backlog finding behind that.

A qualification checklist

The proposals that lose are usually knowable before they are written. Five questions, asked on the first call, sort most of them out.

• Budget. Has the client named a figure, and is it enough for the scope they describe at a fee that covers you? A client who will not discuss money on the first call is telling you how the fee negotiation will go.
• Scope. Can you describe the project in two sentences the client agrees with? If the brief shifts within the call, propose a paid feasibility stage rather than a fixed fee.
• Decision-maker. Is the person you are talking to the person who signs, and is there a partner, parent, board or lender who has not been in the room? Ask directly who else decides.
• Timing. Is there a date that forces a decision, such as a lease, a school year or a planning deadline? A project with no date drifts, and the proposal ages with it.
• Fit. Is this the work you named on your website and want more of? A proposal for work you do not want, at a fee you would not normally accept, costs the same hours as one you do.

Score each answer and set a line below which you do not write a proposal. The enquiry to consultation playbook has the first-call script that gets these answers without an interrogation.

Declining well

A decline that leaves the door open costs an hour and earns referrals. A proposal that loses costs days and earns nothing. Say no to the project rather than to the person.

Keep a list of the firms you refer to. A declined enquiry sent somewhere good comes back as a referral more often than a lost proposal does.

Measuring your own win rate

Most small firms cannot state their win rate, which is why it drifts. The measure is simple: proposals signed divided by proposals sent, over a rolling twelve months, by count and by fee value. Keep three lines alongside it: enquiries received, enquiries declined before a proposal, and proposals still open. The declined line is the one that improves first when qualification tightens, and it should.

• Count by fee value as well as by number. Winning three small jobs and losing one large one is a 75% win rate and a bad year.
• Record the lost reason in one word: price, timing, competitor, went quiet, no decision. Went quiet and no decision are qualification failures rather than pricing failures.
• Compare against the right population. 45% is Deltek’s architecture-only figure for FY2025; 50% was the median across all A&E firms in the 2025 study. A residential sole practitioner working from referrals should expect to beat both, because a referred enquiry is already half qualified.
• Mind the sample. A firm that sends ten proposals a year cannot read a trend from one quarter. Look at twelve months, then at the lost reasons.

One last cause of lost proposals has nothing to do with qualification. We have written about why studios lose projects they have already won: the enquiry that was answered late or not at all. Fix the response before the proposal.