The average company that answers RFPs submits 166 of them a year and spends about 33 hours on each one.
That is roughly 5,500 hours. Three people, full time, all year, producing documents.
Now the number that should stop you. Loopio’s 2026 report, surveying 1,533 people who do this work, puts the 2025 average win rate at 39%, down from 45% the year before. The same period in which 79% of proposal teams started using generative AI to write the things.
Win rates fell six points in the year the tooling arrived.
Nobody did anything wrong. The tooling worked exactly as advertised, the cost of producing a response fell, so everybody produced more of them, and the pool got more crowded than the efficiency gain could cover. The productivity was real and it was competed away inside twelve months.
Which means the useful question is not how to respond faster. It is what you are bidding for, and there are only three honest answers.
The three things you can be bidding for
Most firms bid as though the answer is always the first one. That is why a loss feels like waste rather than like a cost you chose.
To win this deal. Legitimate, and it depends almost entirely on whether you were in the room before the document existed. More on that below.
To be the known second name at the next cycle. Also legitimate, and genuinely valuable in categories where contracts run three or five years, because the incumbent will eventually fail at something and somebody will need a name. But this is a completely different document. Short, memorable, one differentiator stated plainly, no eighty pages of compliance theatre nobody reads. A tenth of the effort, and the goal is that the evaluator remembers your name in two years rather than that you score well now.
To learn how this buyer decides. Worth doing once with an account you want for the long term. You find out who sits on the panel, what they weight, what language they use internally. Once, not annually.
Decide which of the three before anybody opens a template. The failure mode is spending purpose-one effort on a purpose-two opportunity, which is most of the 5,500 hours.
The document tells you which situation you are in
The tell is in the paperwork, not in the relationship, and it takes twenty minutes to read for it.
Specifications written in one vendor’s vocabulary. Not a requirement for a capability, a requirement for a feature by the name a specific product gives it. Somebody pasted that in from a brochure, and it was not yours.
A timeline too short to do the work properly. Three weeks to respond to something that needs a site visit and a discovery call. A buyer who genuinely wants a competitive field gives time for one; a buyer who needs three quotes for a file does not.
No named stakeholder you can speak to before submission. Procurement as a letterbox. If you cannot get a conversation with the person who will use the thing, you are competing on a document against somebody who has had that conversation for two years.
Evaluation criteria that are either missing or oddly precise. Absent weightings mean the decision is discretionary. Weightings to two decimal places sometimes mean a scoring model was built backwards from a preferred answer. Both are worth noticing.
One vendor’s benchmarking estimates that incumbent renewal and expansion RFPs win at something like 60% to 80%, while cold RFPs with no prior contact land at 10% to 20%. That is their own estimate rather than audited data and I could not find an independent source for it, so hold the numbers loosely. The attribution is the part I would trust, because it matches what I have seen: the gap is mostly about whether you helped shape the requirements before they were written.
If you are reading the requirements for the first time, somebody else read them while they were being written.
That is not a conspiracy. It is just what happens when a buyer does their homework by talking to a vendor, which is a reasonable thing for a buyer to do.
Public sector is a different question
Be careful about applying any of this to government and public bodies, because there the tells mean something else.
Procurement rules require a competitive process. So a short timeline, a letterbox contact and a rigid specification are often compliance with the law rather than evidence that the thing is wired. The process is the rule, not a signal.
The questions that matter there are different. Can you meet every mandatory criterion, including the ones about insurance, accreditation and prior contracts of similar size, since failing one of those is a disqualification regardless of how good the rest is. And have you used the pre-bid clarification window, which is usually the only legitimate route to influence a specification and which most bidders ignore entirely.
About the AI claims
While you are reconsidering your process, treat the tooling claims with the same skepticism you would apply to anything else.
The vendors in this space advertise large win-rate improvements, and those figures are self-reported. They also sit awkwardly beside an industry win rate that fell six points in the year their adoption reached four fifths of teams.
Both things can be true at once. Early adopters may well have gained while the average fell. But if the aggregate is down, the honest reading is that this tooling lets you produce more responses rather than better ones, and producing more responses is the thing that stopped working.
Use it to cut the hours on the parts nobody reads. Do not use it to bid on more.
The number to put on the wall
One split, and almost nobody has it.
Separate your win rate for RFPs where you had spoken to somebody at that organization before the document was issued, from your win rate on the ones that arrived cold.
That is the only cut of the data that changes a decision. If you are like most firms, the first number will be high enough to justify serious effort and the second will be low enough that purpose one is not available to you on those bids.
Then put a cost next to it. Hours across everybody who touched the last response, including the subject matter expert who gave up a Thursday and the director who rewrote the executive summary. An RFP response is the most expensive single document most B2B firms produce and the only one whose cost is never counted.
Where I come out
Most RFPs that arrive cold are somebody else’s deal, and that is fine, because there is still a reason to be in some of them.
Just pick the reason first. Bid to be remembered rather than to win, write a tenth as much, and put the hours you saved into the conversations that get you into the next document while it is still being drafted.
That is the whole game, and it happens before the RFP exists.
If you’d like to think through which of your own bids are worth real effort, drop me a line on email, WhatsApp, phone or LinkedIn and we can have a quick chat, and my notes on lead generation cover the part that gets you into the room early. I’m taking on a small number of contracts at the moment, so I’ll tell you straight whether it’s something I could help with.