We set out to publish average proposal turnaround by firm size. The number is quoted constantly, it sounds eminently measurable, and we could not verify a credible public dataset that contains it.
What exists instead is three different measurements, published by three organisations, which get quoted interchangeably. That is worse than an absence, because it produces confident planning assumptions built on the wrong clock.
The market has three clocks
Almost every "proposals take X" claim in circulation collapses these. Once you separate them, most published comparisons stop being comparisons.
The firm-size number is real, but it is effort
Loopio provides the strongest answer to the firm-size part of the question, and it is genuinely citable. Its article published on 26 March 2026 reports response effort of 27 hours for SMBs, 33 hours for mid-market and 39 hours for enterprise organisations.
The methodology is unusually explicit for vendor research: 1,533 response-management professionals across more than 17 industries, developed in collaboration with APMP.
It still does not support "average proposal turnaround by firm size", because it is a workload measure. Quoting those hours as turnaround is the single most common error we found, and it makes larger firms look slower when what the data shows is that they spend more person-hours.
The calendar-time number, and its limits
QorusDocs gets closer to the operating question. Its benchmark, based on nearly 300 proposal respondents, reports in analysis published 23 March 2026 that 51% of respondents say an average request takes 6 to 10 business days, with the professional services segment producing the same result and some responses extending beyond 20 days.
That is elapsed time, and it includes a professional services cut, which is exactly the right shape. What it does not do is segment by firm size. So between the two sources you can have firm size or you can have elapsed time. You cannot have both.
Why the cross-tab does not exist
APMP has done genuine benchmarking, and its US Bid and Proposal Industry Benchmark collects data directly from members. We could not verify a turnaround-by-firm-size result from the publicly accessible material, and we could not establish a publication date for it.
The honest summary is narrow and worth stating precisely: there is no credible public benchmark that simultaneously measures genuine elapsed turnaround, isolates professional services, and segments by firm size. The sloppy version of this article, that no proposal benchmarks exist, would be wrong. Two good ones exist. Neither answers the question.
How to measure it properly
You can establish the missing number internally, and doing so is more valuable than borrowing someone else's. It requires treating proposals as events rather than asking people for a remembered average.
- Timestamp the stages separately: opportunity or RFP receipt, qualification decision, formal kickoff, first complete draft, final approval, submission. APMP's published lifecycle supports treating these as distinct rather than as one undifferentiated cycle.
- Record effort hours separately from elapsed business days. Conflating them is the original error.
- Capture the buyer's window, because a 6-day turnaround inside a 7-day window is a very different operational fact from the same 6 days inside a 30-day window. UK Cabinet Office guidance notes there are mandatory minimum periods in competitive tendering and no set maximums, so windows vary widely by design.
Three or four months of that data will tell you more about your own bottleneck than any industry average, because the useful question is rarely "are we slow" but "which stage absorbs the time".
Why the number gets misquoted
The confusion is not random. It follows a predictable path, and recognising it saves you from repeating it in a board paper.
A vendor publishes a figure with its metric clearly labelled. A journalist or content marketer picks it up and drops the label, because "33 hours per RFP" is less quotable than "RFPs take 33 hours". A third party then compares that against an elapsed-time figure from a different source, concludes that one segment is dramatically slower than another, and publishes the comparison. Two accurate measurements become one wrong claim in three steps, none of which involved anybody lying.
The tell is a comparison that spans sources. If a single chart shows effort by firm size alongside elapsed days by industry, the underlying data almost certainly came from two surveys with different populations, different definitions and different years.
The clock you do not control
There is a third measurement that rarely appears in vendor benchmarks and often dominates the real timeline: the window the buyer gave you.
UK Cabinet Office guidance is explicit that competitive tendering has mandatory minimum time periods and no set maximum. Buyers set windows to suit their own governance, and those windows vary enormously for reasons that have nothing to do with your capacity.
This matters operationally because the same six-day turnaround means opposite things in different windows. Six days inside a seven-day window is a team at full stretch with no slack for a late input. Six days inside a thirty-day window is a team that started late and left review time on the table. An average that does not record the window cannot tell those apart, which is why an industry figure is close to useless for capacity planning even when it is accurate.
APMP's own operating model treats timing as something teams create inside the bid process, setting milestones for writing, review and approval rather than inheriting a single deadline. That is the right frame: the buyer sets the outer bound, and everything inside it is a scheduling decision you own and can measure.
Common questions
So is 6 to 10 days a fair planning assumption?
As a rough elapsed-time expectation for professional services responses, it is the best public figure we found. Treat it as a distribution rather than a target, given the same source reports responses extending past 20 days.
Why do vendors publish these numbers at all?
Because faster turnaround is the product benefit they sell. That does not make the data wrong, and the Loopio methodology in particular is more transparent than most. It does mean the framing around the number is chosen to favour the vendor.
Does faster turnaround actually win more?
We found no credible evidence linking elapsed turnaround to win rate in professional services. It is plausible, widely assumed, and unproven. Speed has obvious internal value in capacity terms regardless.
What is a realistic first target if we start measuring?
Do not set one. Measure for a quarter first, because the common discovery is that the drafting is not the bottleneck. Waiting for internal approvals and waiting for input from the deal team usually dominate, and neither is fixed by writing faster.