SOVE7GN GTM
How to run a discovery call: a 30-minute agenda and the questions to ask
Updated 14 November 2026 · Written by Reyansh Juneja and Naman Dhiman, the co-founders of SOVE7GN, who run inbound, midbound and outbound for founders
A discovery call is 30 minutes spent finding out whether this buyer has a problem worth paying to fix, and whether you are the one who should fix it. Run it in five blocks: 2 minutes to set the agenda, 5 minutes on their situation, 12 minutes on the problem and what it costs them, 6 minutes on how a decision gets made there, and 5 minutes to book a dated next step or agree that it is a no.
Short answer: ask around a dozen targeted questions, let the buyer talk more than you, and never end without a date. Gong's 2025 analysis of 326,000 sales calls of 10 minutes or more found sellers talked 57% of the time in won deals and 62% in lost ones, and that won deals averaged 15 to 16 questions while lost deals averaged about 20 (Gong Labs, checked November 2026).
What does a 30-minute discovery call look like, minute by minute?
The agenda we would run as a founder selling a product worth roughly $25K a year, assuming five minutes of homework and a video call.
| Minutes | What you are doing | What you say |
|---|---|---|
| 0 to 2 | Set the frame | "I have 30 minutes. I want to understand how you handle this today, and if it looks like a fit I will show you how we do it. If it is not a fit I will say so. Does that work, and is there anything you want to make sure we cover?" |
| 2 to 7 | Their situation today | How the job gets done now, who does it, which tools, since when. Facts only, no selling yet. |
| 7 to 19 | The problem and what it costs | What breaks, how often, what it costs in money, time or churn, what they already tried, what happens if nothing changes. This block decides the deal. |
| 19 to 25 | How a decision gets made there | Who else cares, what they bought like this before, what the approval path is, what the timing pressure is. |
| 25 to 30 | Next step or a clean no | Summarise what you heard, say whether you can help, then either book the next meeting during the call or end it honestly. |
The problem block is the longest for a reason. Gong's analysis of more than 519,000 B2B sales calls reports that the best calls surface three to four business problems rather than one or ten, and that top sellers spread questions across the call instead of front-loading them (Gong Labs, checked November 2026). Gong calls its explanation for that figure a hypothesis, so treat it as a shape, not a target.
What questions should I ask on a discovery call?
Ten is enough for 30 minutes. These are the ones that move a deal, with what a useful answer sounds like and what a dead end sounds like.
| Question | A good answer sounds like | A bad answer sounds like |
|---|---|---|
| "Walk me through how this gets done today." | A named person, a named tool, a weekly rhythm | "It depends" with no example |
| "What made you take this call now?" | A trigger: a hire, a target, a churned account, a board question | "We are just looking around" |
| "How often does that go wrong, and what does it cost when it does?" | A number they already track, even a rough one | "It is hard to quantify", twice in a row |
| "What have you already tried?" | A tool they bought, a person they hired, a process that failed | Nothing, ever |
| "Who else feels this problem?" | Two or three names and roles | Only the person on the call |
| "If you do nothing for six months, what happens?" | A consequence with a date attached | "Nothing much, it is annoying" |
| "What would have to be true for you to buy something like this?" | Criteria, even rough ones | "I would need to think about it" |
| "When you bought something similar, how did that go?" | A real process with names and steps | "We have never bought anything like this" |
| "What is your timing driven by?" | A quarter, a renewal, a launch, a board meeting | "No rush" |
| "What would stop this from happening?" | A named blocker they will go and handle | Silence |
The right-hand column is what founders talk past. "No rush" plus "it is annoying" plus one person who cares is not a deal this quarter, whatever the enthusiasm sounded like.
How much should I talk on a discovery call?
Less than you want to. Gong's original 2016 figure for the ideal split was 43% talking and 57% listening, and its 2025 refresh of 326,000 calls found sellers talking 57% of the time in won deals against 62% in lost ones (Gong Labs, checked November 2026). Gong does not reconcile those two numbers, so do not chase a percentage. The direction holds across both: in deals that are lost, the seller talks more.
The same 2025 piece reports that high performers keep roughly the same ratio whether they win or lose, while low performers talk more as a deal goes badly. If you catch yourself explaining more in the second half of a call, the call is already going the wrong way, and more explaining will not fix it. Ask another question instead.
On question count, the 2025 data says won deals averaged 15 to 16 questions and lost deals about 20, while the older Gong analysis puts the sweet spot at 11 to 14 targeted questions. Call it a dozen.
How do I end a discovery call so the deal actually moves?
Three rules, in this order.
- Say what you heard, out loud. Three sentences: the problem, what it costs, what they want instead. If you got it wrong they will correct you, and that correction is worth more than the rest of the call.
- Say whether you can help, including when you cannot. "We are not the right fit, here is what I would do instead" costs you nothing and it comes back later.
- Book the date during the call. Open the calendar, agree a time, send the invite before you hang up. "I will follow up next week" is how a good call becomes a dead one.
Then send a recap the same hour: the problem in their words, the next step, the date. Speed matters after a call for the same reason it matters before one, which is the argument on our page about speed to lead.
When should I disqualify someone on the call?
Say no, on the call, when at least two of these are true: nobody has tried to fix this before, no one else in the company feels it, no date is forcing a decision, and they cannot describe what the problem costs in any unit at all. A seed-stage founder has maybe 15 first calls in a good month, which is the arithmetic behind how many sales calls a startup needs. Spending three of them on somebody who is reading the market is a quarter you do not get back. Which framework to record all this in is a separate question, and the frameworks founders reach for (BANT, MEDDIC, CHAMP) were built for different deal sizes: we compared them in our guide to the lead qualification framework.
Where do these calls come from in the first place?
A good agenda does not help if the calendar is empty. That is what SOVE7GN GTM is built for: we run the founder's daily posting, the fast follow-up with people who show interest, and the cold outreach, and we book the calls. You take them. Book 30 minutes to see how that would work for your company, and pricing is shared on that call. If you are keeping all of it in-house for now, the trade-offs are in founder-led sales, and the rest of these guides are in the GTM guides hub.
Questions people ask
How long should a discovery call be?
Thirty minutes is enough for a first call with most B2B buyers, and it is easier to get accepted than an hour. Keep the hour for the second call, when more people are in the room and you are showing something.
What questions should I ask on a discovery call?
Ten to a dozen targeted questions covering how the job is done today, what triggered the call, what the problem costs, what they already tried, who else cares, and how they buy. Gong's data puts the useful range at 11 to 14 questions in one analysis, and 15 to 16 in won deals in another.
Should I demo the product on a discovery call?
Only a short, targeted piece, and only once you know which problem you are aiming at. A full demo in the first 30 minutes replaces the questions that decide whether a deal exists.
What do I do when the buyer will not give me a number for what the problem costs?
Offer a range and let them correct you: "is this a few hours a week or a few days?" People who will not estimate at all usually do not own the problem, which is a qualification answer in itself.
Want calls on your calendar without hiring five people?
Book a 30-minute call. Tell us who buys from you and we'll show you what inbound, midbound and outbound would look like for your startup.
Book a call How SOVE7GN GTM works- Lead qualification frameworkBANT, CHAMP, MEDDIC and MEDDPICC compared, and what to use at seed.
- Founder-led sales: keep or hand offWhat a founder keeps, what to hand off first, and a handoff checklist.
- How many sales calls per monthWork back from ACV and close rate to the calls you need each month.
- When a prospect says send me more infoWhy prospects say it, and word-for-word replies for email, calls and LinkedIn.