The call felt good. The prospect was friendly, the AE got through the list, and the notes say "interested, pain around reporting, follow up in two weeks." Three weeks later the opportunity sits in stage two with no next meeting, and nobody can say whether it was ever real. The manager listens to the recording and hears 22 questions, a ten-minute product tour in the middle, and no moment where anyone asked what the reporting problem costs.
That is the usual failure of a discovery call script. It tells the rep what to ask, not what to find out, and it ends without a verdict. This framework fixes both. Before the call, the rep writes pain hypotheses. During the call, each stage has a time budget and a job. After the call, six criteria produce a score the CRM can store and a manager can check.
Taking a discovery call this week? Pitch Pacer turns a 30-minute discovery call into a timed, stage-by-stage script: tell it who is on the call and what is broken, and it gives you minutes for each stage. The free tool opens with a work email.
Those three numbers describe the pressure on every first call. Research by Adience, reported in December 2025 from a survey of 350 business buyers across four regions, found that 30 percent said vendors asked poor or repetitive discovery questions, and 29 percent said vendors did not understand their industry or use case. Gartner's June 2025 survey of 632 B2B buyers found that 61 percent prefer an overall rep-free buying experience. Buyers who would rather not talk to you will not sit through an interview your website could have answered.
Gong's analysis of 519,000 recorded discovery calls, first published in 2017 and updated in 2026, found that asking between 11 and 14 questions correlated with the greatest success, and that the most effective calls uncovered three or four business problems. A more recent Gong Labs study of 326,000 sales calls found that won deals averaged 15 to 16 questions per call, against about 20 in lost deals. More questions is not more discovery. Better-aimed questions are.
That makes discovery a sales operations problem, not a talent problem. Strong reps do this by instinct; the script is how the rest of the team does it on purpose.
Diagnosis: why most discovery call scripts produce friendly calls and empty pipeline
Scripts fail in predictable ways. Each of the four below shows up in the recording and in the CRM.
1. The script is a questionnaire, not a plan
The common discovery call script is a list of 20 to 40 open questions grouped by topic: current process, tools, team, budget, timeline. A rep working through it in order produces exactly what the Adience respondents complained about, a call that feels repetitive because half the answers were on the buyer's LinkedIn page or your own intake form. A questionnaire also has no priority, so when time runs short the rep skips the last questions, usually the ones about money and decision process.
2. There is no hypothesis, so the pain stays generic
"What are your biggest challenges?" invites a polite, general answer. A rep who arrives with a specific guess ("teams at your stage usually find that reps stop updating close dates after the second call, so the forecast drifts; is that happening here?") gets a specific yes, no or correction. Each of those is useful. The Gong finding that the best calls surface three or four business problems is easier to reach when the rep brings candidate problems to confirm, rather than hoping the buyer volunteers them.
3. The rep starts pitching before the problem is sized
In Gong's study of 326,000 calls, won deals averaged 57 percent rep talk time and lost deals 62 percent, and low performers' talk time rose from 54 percent in won deals to 64 percent in lost ones. The usual trigger is the first hint of pain: the buyer mentions reporting, and the rep shares a screen. The problem never gets a number, so the later proposal has nothing to be measured against.
4. The call ends without a verdict
Most scripts end at "next steps," which in practice means "I'll send some materials." Nobody records whether the call found a real problem, whether it matters now, or who decides. A month later, pipeline review is an argument about feelings, because the evidence was never captured in a form anyone can compare.
Framework: the Hypothesis-to-Score discovery call script
The framework has three parts that map to before, during and after the call. Everything below is a suggested starting point for a 30-minute first discovery call at a $3M to $30M ARR B2B SaaS company, not a benchmark. Adjust it to your sales cycle, and write the changes down.
Before the call: write two or three pain hypotheses. A pain hypothesis has three parts: the problem you suspect, the evidence that made you suspect it, and the cost if it is true. "They have grown from 8 to 20 AEs in a year (LinkedIn); teams that scale that fast usually lose track of handoffs between SDR and AE; if so, a share of booked meetings never get worked." Each hypothesis gets one confirming question and one sizing question. That is six of your 11 to 14 questions, all specific.
During the call: six stages with a time budget.
| Stage | Minutes | The job of this stage | Done when |
|---|---|---|---|
| 1. Open and agree the agenda | 2 | Confirm time, state what you want to learn, ask what they want to leave with | The buyer has added or confirmed one goal for the call |
| 2. Confirm context | 3 | Check two or three facts you already researched, not ask for them | Your account notes are corrected, not rebuilt |
| 3. Test the hypotheses | 11 | Put each pain hypothesis to the buyer; follow the ones that land | Each hypothesis is confirmed, rejected or replaced |
| 4. Size the impact | 6 | Turn the confirmed problem into a number: hours, deals, dollars, risk | The buyer, not the rep, has said what it costs |
| 5. Map the decision | 5 | Who else is affected, who decides, what has to happen before a purchase | You can name the next person to talk to |
| 6. Agree the next step | 3 | Propose a dated next step tied to what you learned | A meeting is on the calendar, or both sides agree to stop |
Two questions decide whether the call was worth having, in our framing: is there a problem worth solving now, and is there a path to a decision? Stages 3 and 4 answer the first. Stages 5 and 6 answer the second. If stage 3 rejects every hypothesis, the honest move is to spend the remaining minutes learning why, not to pitch.
Notice what is missing: a product demo. If the buyer asks, show the one screen that relates to the confirmed problem, then return to sizing.
After the call: score it on six criteria, 0 to 2 each.
| Criterion | 0 | 1 | 2 |
|---|---|---|---|
| Pain confirmed | No hypothesis landed | Pain acknowledged in general terms | Buyer described a specific instance in their own words |
| Impact sized | No number | Rep's estimate, not confirmed | Buyer stated or agreed a cost |
| Urgency | "Someday" | A vague target date | A dated event drives it: board meeting, renewal, hiring plan |
| Decision path | Unknown | Decision maker named | Decision maker and steps to purchase described |
| Fit | Outside your ideal customer profile | Partial fit | Clear fit on segment and use case |
| Next step | "Send me something" | Next step agreed, no date | Dated meeting with the right people |
A total of 9 to 12 advances the opportunity. A total of 5 to 8 means a second discovery call with a specific gap to close, usually impact or decision path. Below 5, the lead goes back to nurture with a note on why. These cut-offs are a suggested starting point; after a quarter, check them against which scored calls actually turned into won deals and move them.
Implementation: rolling the script out across an AE and SDR team
Expect three to four weeks from first draft to a scored call becoming routine.
Pull ten won and ten lost discovery calls from the last two quarters. Listen for which problems came up in won deals, how early the price of the problem was discussed, and how each call ended. This seeds the hypothesis library and shows which criteria separate wins from losses.
Write a hypothesis library by segment and persona. Three to five hypotheses per segment, each with its evidence signals, a confirming question and a sizing question. Keep it to one page per segment so reps use it. A CFO, a VP of Operations and a CTO feel the same problem differently, so write the sizing question in each persona's units: dollars, hours or risk.
Add the six score fields to the opportunity record. One picklist per criterion (0, 1, 2), a calculated total, and a required free-text field for the strongest piece of evidence. Make the fields required before the opportunity can leave the discovery stage.
Run the script on live calls for two weeks with manager review. Managers score the same recordings independently. Where the manager and rep scores differ by more than two points, the definitions in the rubric need tightening, not the rep.
Make prep part of the calendar. Block ten minutes before each discovery call to write the hypotheses and set the time budget. A tool like Pitch Pacer, or a shared template, makes that a habit rather than an intention.
Recalibrate after one quarter. Compare discovery scores to outcomes. If calls scoring 9 or more close at a rate no better than calls scoring 6, one of the criteria is not predictive; find it and replace it.
Workflow: who owns each part of the discovery call
Research the account, choose two or three hypotheses from the library, adapt them with evidence specific to this buyer, and set the time budget. If an SDR booked it, their notes are the first input.
Run the six stages, keep to the minutes, and score the call within the hour while it is fresh. Write the strongest piece of evidence in the buyer's own words.
Review two scored calls per rep per week against the recording. Coach on the gap between rep score and manager score, not on style.
Owns the score fields, the stage gate, the hypothesis library and the quarterly recalibration. Reports scores by rep and segment next to conversion to the next stage.
Board narrative: explaining discovery quality in three sentences
Boards ask why win rates fell or the forecast slipped. These three statements connect the answer to the first call.
"Every first call now starts from written hypotheses about the buyer's problem and ends with a score out of 12 stored in the CRM, so we qualify on evidence, not on how the call felt."
"We compare discovery scores to outcomes each quarter. Opportunities that enter pipeline with a sized problem and a named decision path should convert at a visibly higher rate, and if they do not, we change the criteria."
"Buyers are less patient with sellers than they were. Gartner found in 2025 that 61 percent of B2B buyers prefer a rep-free experience, so the calls we do get have to earn the next one."
Cross-domain: where discovery connects to the rest of the revenue engine
A discovery score is only useful if it travels. The evidence captured on the first call should reach whoever owns the opportunity next, which is the job of a Handoff Orchestrator: the AE who inherits an SDR-booked meeting, and later the CSM who inherits the account, should see the confirmed pain and its sized impact without asking the buyer again. Our piece on why deals die between marketing, SDR and AE shows where that context usually drops.
Downstream, the score feeds pipeline quality. A Pipeline Hygiene Sentinel can flag opportunities that advanced past discovery with a low score or an empty evidence field, and a forecast is easier to trust when every deal in it entered with a sized problem. Upstream, discovery scores are the ground truth for stage definitions; our guide to evidence-gated lifecycle stage logic uses the same principle for the MQL-to-SQL boundary. If you are weighing a conversation intelligence platform to record and analyze these calls, our comparison of Gong, Clari and Salesforce Einstein covers what each one actually fixes.
Frequently asked questions
What should a discovery call script include?
A discovery call script should include two or three written hypotheses about the buyer's problem, a time budget for each stage of the call, a sizing question for each hypothesis, and a way to record the result. The structure matters more than the questions: a plain question list leaves the rep without priorities and ends without a verdict.
How long should a discovery call be?
Thirty minutes is enough for a first discovery call with one or two buyers if the rep has prepared hypotheses and does not demo the product. Longer calls usually mean the rep is collecting facts research could have answered, or pitching. Complex deals with several stakeholders may need a second discovery call, which is better than stretching the first one past the buyer's attention.
How many questions should you ask on a discovery call?
Gong's analysis of 519,000 discovery calls found that 11 to 14 questions correlated with the best outcomes, and a later Gong study found won deals averaged 15 to 16 questions per call against about 20 in lost deals. Aim for fewer, better-aimed questions: one confirming and one sizing question per hypothesis, plus a few on the decision process.
What are the stages of a discovery call?
A practical structure has six stages: open and agree the agenda, confirm the context you researched, test your pain hypotheses, size the impact of the confirmed problem, map the decision process, and agree a dated next step. In a 30-minute call, most of the time belongs to testing hypotheses and sizing impact, because that is where the opportunity is won or lost.
How do you score a discovery call?
Score it right after the call on a small number of criteria, each with a clear definition. A simple version rates pain confirmed, impact sized, urgency, decision path, fit and next step from 0 to 2, for a total out of 12. Store each criterion as its own CRM field so you can compare scores to outcomes each quarter and adjust the criteria that do not predict wins.
Sources: Adience survey reported by eCommerceNews, December 2025 (350 business buyers; fieldwork dates not published). Gartner, June 2025 (632 B2B buyers surveyed August to September 2024). Gong Labs discovery study (519,000 calls, published July 2017, modified March 2026). Gong Labs talk-to-listen study (326,000 calls, published August 2025, modified March 2026). VANDFORT Pitch Pacer page (accessed November 2026).
