Discovery Call Questions Template
Four sheets and four documents that score each call from what the prospect said, then prune the question set on your own closed deals.
Free download · No account needed
Question Effectiveness, quarterly review
13 questions, 113 closed deals, minimum sample 30
| Deals in the analysis | 113 closed, won and lost both |
| Questions with a verdict | 11 |
| Too few to judge | 2 |
| Cut this quarter | 2 |
1. Kept, largest separation first
| Question | Asked | Win asked | Win not | Diff |
|---|---|---|---|---|
| Why now, rather than six months ago | 74 | 44% | 18% | +26 |
| What happens when it goes wrong | 88 | 41% | 24% | +17 |
| What does success look like in 12 months | 49 | 42% | 26% | +16 |
2. Cut on evidence, 30 September
| Question | Asked | Diff | Why |
|---|---|---|---|
| On a scale of one to ten, how painful is this | 61 | -1 | Answered every time, never varied |
| Are you the decision maker | 84 | -4 | Yes on 84% of calls, including the stalls |
3. Too few to judge is a verdict, not a blank
Two questions show promising separation on 7 and 11 deals. Both read too few to judge with the shortfall stated. A win rate computed on seven deals gets quoted in a pipeline review and never retracted.
The 12-month success question is asked almost only by the two most senior reps, who close more anyway. The confound is recorded beside the number.
Every sales team already has a list of discovery questions, and almost none of them knows which questions on the list matter. Search the phrase and page one hands you thirty questions grouped by framework plus a scorecard with point values printed on it: twenty-five for approved budget, fifteen for planned, disqualify below forty. Nobody calibrated those numbers against your deal size, your sales cycle or your buyers. They are somebody else's arithmetic, adopted whole. Separately, whether your own scorecard reads the same thing twice is measurable.
The number that would actually help is one nobody computes: the win rate on deals where a question was asked against deals where it was not. That needs the transcript and the outcome in the same place, so it needs recordings, which is why the log here carries a consent basis column. Federal law permits a party to the call to record it. California requires every party to consent. A call with no basis recorded is excluded.
So the scorecard has four answer states rather than a tick: quoted, paraphrased, inferred and not asked. Inferred becomes an open item rather than a score, because budget inferred from headcount is a guess wearing a number. Question Effectiveness then reports separation with the sample size beside every figure, and writes too few to judge where the sample is short. The gap register splits what nobody asked into ran out of time, deflected, and never asked, because those three go to different people.
What's in the pack
Question Set by Persona
Every question carries what it establishes, what a good answer contains, the one follow-up for a thin answer, and its current verdict from the effectiveness sheet.
Question Effectiveness
Win rate on deals where each question was asked against deals where it was not, with the sample beside every figure and a confound column for the obvious explanations.
Qualification Scorecard
One row per element attempted, with four answer states instead of a tick and the prospect's own words in the evidence column.
Discovery Gap Register
What the team keeps failing to ask, split into ran out of time, asked and deflected, and never asked, ranked by the value of the open deals affected.
Call to Outcome Log
One row per call joining it to its deal, its persona, its consent basis and its eventual outcome. This is the join the effectiveness analysis runs on.
Call Structure
The shape of a first call rather than a script, with the current-state questions before the problem questions and the why-now question placed before the overrun.
Disqualification Standard
The five reasons to stop working a deal, each confirmed from what the prospect said, with a threshold derived from the point below which nothing has ever closed.
Follow-up Summary Format
Two documents from one call, kept separate: the forwardable version for the prospect and the CRM version with commitments quoted verbatim.
How to use it
- 1
Open in River, or take it blank
Open the pack in River and hand it your transcripts, or download the Word documents and CSV sheets and fill them in yourself.
- 2
Send calls with outcomes
Transcripts from Gong, Chorus, Zoom or a pasted export, plus what happened to each deal. Old calls beat recent ones because their outcomes are known.
- 3
Score from what they said
Each element gets an answer state and a quote. Inferred rows become open items with owners rather than filled cells nobody trusts.
- 4
Prune the set, close the gaps
Keep, rewrite, cut or too few to judge on every question, then the gap register by value at risk with a named owner for each fix.
Frequently asked questions
Is this template free?
Yes. Four Word documents and four CSV sheets, downloaded as a zip, no signup. River is the optional half: it scores your past calls from the transcripts, computes the separation, and writes the gap register. Neighbouring packs sit under the template library, and one-shot jobs under the tool index.
How many calls do we need before the numbers mean anything?
Enough closed deals, not enough calls. The sheet carries a minimum sample and reports too few to judge below it, so with a dozen transcripts you get a scored question set and no effectiveness verdicts. Thirty closed deals per question is where the numbers start meaning something, and closed-lost counts.
We use MEDDIC. Does this assume a framework?
Whichever you already use. The question set is grouped by persona and each question names the element it establishes, so it maps onto BANT, MEDDIC or SPICED without rewriting anything. What the pack will not hand you is the framework's point values, because those come from your own closed records or they do not come at all.
Do we have to record calls for this to work?
No, and calls without a recording are still scored from notes. They are excluded from the effectiveness analysis, and the log says so in its own column rather than silently dropping them. Where you do record, the consent basis is recorded too: one party under federal law, every party in states that require it.
What stops a question surviving because a senior person likes it?
Evidence and a stated sample. A question with a large sample and no separation is cut, and the sheet records what it was cut on and the date so nobody reintroduces it in January. A question with real separation but only seven deals behind it stays as too few to judge, not as a finding.
Does it help with one specific upcoming call?
The prepare prompt writes a one-page brief: three hypotheses to test, the questions in order, the elements already answered so nobody asks twice, and what to drop if the call runs short. It works from whatever you know, and an account research brief sharpens it. Afterwards, the transcript becomes CRM field changes rather than a note, and it sets the minute budget for the demo.
How does it handle a question only the best reps ask?
It names the problem. A question asked almost only by your two senior reps will show a high win rate, because senior reps close more whatever they ask. The sheet carries a confound column that says exactly that, rather than reporting the lift and letting somebody roll it out. Spread the question first, then measure it.
Find out which of your questions predicted a win
Take the Word documents and CSV sheets blank, or open this exact pack in River and let it score your finished deals first.
Edit with AI