Count a spoken yes only after you email the exact terms and they reply to confirm. If no reply comes, you have learned something early.
The buyer said "we're going ahead" on the call. Three weeks later, nothing is signed and nobody remembers the same terms.
Treat a spoken yes as nothing yet. Send one short message that spells out the exact terms you believe were agreed, numbers included, and ask them to reply to confirm. The deal exists when the reply arrives, not when the words were said.
The reply is also the test. If it does not come, you have learned now, not at the end of the quarter, that the yes was softer than it sounded, was not theirs to give, or meant different terms to them. A missing reply is information, not a reason to chase.
Y Combinator's handshake deal protocol. Written for founders and investors, it is the only published wording of the confirmation. Three steps. The investor says "I'm in." The founder sends an email or text: "Can you confirm you're in for [offer]?", with the exact amount and terms. The investor replies yes. Until all three have happened there is no deal. The page rules out a range, such as "$50k to $150k", which it calls an offer at the bottom of the range plus an expression of interest, and any condition, such as investing only if someone else leads. It suggests both sides send and answer the message on their phones before they part, and treat unwillingness to do so as a warning sign.
Rob Fitzpatrick's The Mom Test gives the test for whether a yes means anything. A compliment costs nothing, so it tells you nothing. What counts is what the other side gives up: time, such as a next meeting with a clear purpose; reputation, such as an introduction to their boss or a public case study; or money, such as a letter of intent or a deposit. His rule of thumb: if you do not know what happens next after a sales meeting, the meeting was pointless.
No B2B sales source publishes the confirmation message itself, so the wording you can copy is the investor one above.
Nobody has measured it. No study compares sellers who put a verbal yes in writing and waited for a reply with sellers who did not.
The nearest measurement is whether buyers write back at all. Gong, which sells the software that counts the emails, reported in October 2019 from more than 500,000 emails in live deals that won deals averaged about eight emails a week each way against just under two in deals never signed, and that the gap widened when counting only emails the buyer started. It is a correlation, since a deal going well produces replies, but it is the only sales data on reading a buyer's silence.
When talk carries information. Joseph Farrell and Matthew Rabin's review of "cheap talk", in 1996, argues that a statement is believable when the speaker would only say it if it were true, and when saying it gives them a reason to follow through. Where the two sides want different things, talk says less. That is the case for making the yes specific and written, not for distrusting it.
People who write down what they will do mostly do it. In Gary Charness and Martin Dufwenberg's experiments at the University of California, Santa Barbara, published in 2006, the player who had to decide whether to honour the other's trust could first send a written message. Of those who wrote a promise, 79% kept it, against 33% of those who wrote something else or nothing. The players chose whether to promise, so the people who meant to follow through were also the ones who wrote it down, which is exactly the ambiguity a seller faces when a confirmation arrives.
The economics evidence comes from students playing for small sums, whose promises were volunteered rather than requested. Nobody has compared deals where the seller confirmed the verbal yes in writing with deals where they did not. The move is better supported as a way of finding out early than as a way of closing.
Send one message with the exact terms, ask for a reply, and count the deal only when the reply lands. If it does not come, you have your answer early. Whether asking for that reply changes the outcome of a B2B deal has never been tested.