Find the one task the buyer still owns before signing, complete as much of it as you can, and hand it over finished.
The buyer agreed to the next step. Two weeks later, the pilot criteria they promised to write are still not written.
Most stalled deals sit on a task the buyer owns and has not done: the internal summary, the vendor form, the pilot criteria. Find that task, do every part of it that does not need their authority, and hand it back finished. Kyle Asay, who ran mid-market sales at LaunchDarkly, puts the reason in his post on unsticking deals: 99% of the buyer's job has nothing to do with buying your product, and 100% of yours is helping them buy it.
Asay's proof-of-concept draft. Read the discovery and demo transcripts and draft the proof-of-concept requirements and success criteria for the buyer: the capabilities they said were essential, and the ones competitors struggle with. Send it with a short note. You have done many of these. This is a starting point built from what they told you. They should add or remove anything. It should save them time.
His two questions for any stuck deal. What is the next step the buyer has to take? What part of it can you do for them?
The reminder that won a megastudy. Two texts, three days apart, telling the person the thing is waiting for them.
Nobody has measured it. No study compares deals where the seller did the buyer's task with deals where the seller waited for it. Asay publishes the move as practice.
The argument for it. Nicholas Toman, Brent Adamson and Cristina Gomez argued in Harvard Business Review in 2017 that buyers are overwhelmed, not short of information, and that the sellers who win clear obstacles out of the buyer's way instead of sending more material. They report that the number of people involved in a B2B solution purchase grew from an average of 5.4 to 6.8 in two years. Their survey figures on giving a clear recommendation are under Offer one recommendation, not a menu.
Students who said yes and then did not turn up. Ben Castleman and Lindsay Page ran a field experiment on summer melt, published in 2015: school leavers who had accepted a college place and then never enrolled. Texts that named the specific tasks still outstanding, sent by the student's own school district or advising organisation with a counsellor answering replies, raised enrolment. The gains went to students who had the least access to college counselling during the year. At the one site where students already had plenty of support, the texts did nothing.
The wording that won a megastudy. Katherine Milkman and colleagues tested 22 reminder texts on 689,693 Walmart pharmacy customers. On average the reminders raised flu vaccination by 2.0 percentage points. The best performer was two texts three days apart telling people a vaccine was waiting for them. Neither experts nor ordinary people predicted it would win.
People want to finish what they started. Maria Ovsiankina's 1928 experiments found that people go back, unprompted, to a task they were interrupted in. That is the reason to hand back a half-finished document rather than a fresh request.
Nobody has tested the move in a deal. The trials that measured the mechanism had individuals finishing their own short tasks: enrolling, getting a flu shot. The evidence covers finishing a task, not making a decision that belongs to someone else.
Find the task the buyer still owes, do the part that does not need their authority, and hand it back finished, with the task named exactly. In the one field trial on people who committed and then stalled, the texts that worked named the specific task and came from the student's own school, with a person answering replies.