B2B discovery call questions: the twelve that decide whether a deal is real

Most B2B deals are lost in the first thirty minutes of the first real conversation, and the seller almost never notices. The call feels warm, the prospect is polite, a follow-up gets promised. Three weeks later the thread goes quiet, because nobody established what the problem was costing or who was allowed to fix it.

Running outbound through Sail Force Marketing, I watch good operators treat discovery as a friendly interview and then wonder why their pipeline is full of deals that never move. The fix is not charisma. It is structure.

What a discovery call is actually for

A discovery call has one job: to produce a mutual decision about whether there is a problem worth paying to solve, and whether you are credibly the one to solve it. It is not a demo, a rapport exercise, or a guided tour of your feature list.

That framing changes what a good call looks like. If the honest answer is no, the call ends in a clean no inside thirty minutes and you both get your week back. I treat selling as stewardship rather than persuasion, and stewardship means I do not spend a buyer's time building a case for something they do not need.

A discovery call is not where you convince someone. It is where you find out whether convincing is even warranted.

The qualification that happens before this call is a separate discipline, covered in the qualification bar that decides whether your pipeline is real. Assume here that the meeting was earned. The question is what you do with it.

Two ratios that decide the call before the content does

Gong analysed over 519,000 B2B sales calls and found sellers have the best shot at a successful discovery call when they ask between eleven and fourteen targeted questions. Fewer and the picture is too thin to qualify on. More and it stops being a conversation. The same research found strong performers spread questions across the whole call, while average performers front-load them like a checklist and then talk for the rest of the hour.

The second ratio is talk time. In a separate analysis of 326,000 calls, Gong found closed-won deals averaged 57% seller talk time against 62% in lost deals, with the long-standing benchmark near 43% talking and 57% listening. Five points is three extra minutes of you in a sixty minute call, almost always spent answering a question nobody asked. Both numbers are measurable on your own recordings this week.

The twelve questions, in the order I ask them

Order matters more than wording. Situation first, because people answer factual questions easily and warm up doing it. Cost second, because you cannot price a solution to a problem nobody has sized. Decision mechanics last, because by then you have earned the right to ask how money gets approved.

Questions one to four: the situation in their words

1. Walk me through how this works today, step by step. Tests whether they can describe their own process. If they cannot, they do not own it.
2. When did this become something you had to deal with rather than live with? Tests for a trigger event. Problems without triggers do not get funded.
3. What have you already tried? Tells you which objections are already dead. A buyer who has tried nothing has a nuisance, not a priority.
4. Who else is affected besides you? Starts mapping the buying group before you have to ask about it directly.

Questions five to eight: the cost of doing nothing

5. If nothing changes for two quarters, what happens? Tests urgency honestly. Most answers are "we cope", which is useful information, not a failure.
6. What is this costing you, in hours or in revenue? Tests whether they can build a business case without you. If they cannot produce a number, you will have to.
7. How did you arrive at that figure? Tests whether the number survives their own finance team. A figure the buyer cannot defend internally gets halved in the approval meeting.
8. What does solving this unlock that you cannot do today? Tests upside. Cost avoidance gets budget approved slowly. Growth gets it approved quickly.

Questions nine to twelve: how the decision actually gets made

9. Last time you bought in this category, what did the process look like? People describe past processes accurately and future ones optimistically.
10. Who signs, and who can stop it? The two roles that matter. They are rarely the same person, and the second is rarely on your first call.
11. What has to be true for this to be approved by the end of the quarter? Tests the timeline against their conditions rather than yours.
12. What would make you decide not to do this at all? Surfaces more real objections than any other question, because it gives permission to say no without ending the conversation.

Twelve fits inside the eleven to fourteen band. If a call only supports eight, the deal is probably not real, which is a finding worth having.

What to listen for while they answer

The questions are the easy half. The value sits in what you do with the answers. Five signals are worth more than the rest of the transcript.

• Numbers without sources. When a buyer says "it costs us about two days a week", ask how they know. Unsourced numbers collapse in the approval meeting you will not be in.
• Pronoun shifts. "I need this" and "they want this" are different deals. The second means your contact is carrying someone else's priority and will drop it under pressure.
• Named people. Every name mentioned is a stakeholder. Write each down with the role attached, and read the list back before the call ends.
• Dates that belong to the business. A board meeting, an audit, a renewal, a funding round. Dates the buyer did not invent for your benefit are the only real deadlines in a deal.
• The second problem. Buyers often mention a bigger issue in passing. Follow it. The larger problem usually has the budget attached.

If you leave a discovery call without a number, a name and a date, you did not run discovery. You had a conversation.

How to close the call so the deal moves

The last five minutes carry disproportionate weight. I do three things in them, in this order.

First, I summarise back in their language: "So the situation is X, it is costing roughly Y, and the reason it matters now is Z. Have I got that right?" Buyers correct summaries readily, and every correction is free qualification.

Second, I ask for the no: "Based on that, is this worth continuing, or should we leave it?" It is the cheapest filter in the process, and buyers respect it because almost nobody offers them the exit.

Third, I book the next step live, with the people named and the calendar open: "The next useful conversation includes whoever owns the budget line. Can we find thirty minutes with them in the next two weeks?" Next steps agreed by email afterwards happen far less often than next steps booked while both parties are in the room.

Five ways discovery calls fail, and what to do instead

• Demoing on discovery. The call becomes a feature review with no qualification in it. Instead, answer briefly and redirect: "I can show you that properly next week once I know which part matters to you. Can I ask two more things first?"
• Accepting vague pain. "Efficiency" and "visibility" are categories, not problems. Instead, ask for the last specific instance: "When did that last cause a real issue?"
• Single-threading. One friendly contact who loves you and cannot sign is the most common shape of a stalled deal. Instead, ask question ten early enough that you still have call time to act on the answer.
• Taking a budget answer at face value. "We have budget" often means someone else does. Instead, ask where the money would come from and what it is currently earmarked for.
• Selling into a no. Pushing costs you the referral as well as the deal. Instead, end cleanly and ask who else in their network has the problem. Clean exits generate more introductions than reluctant maybes ever do.

The four numbers to watch, and where to start

You do not need a new CRM for this. You need four figures reviewed monthly against your own baseline, not an industry average measured on a different business.

1. Questions asked per call. Count them from a recording. If the median is under eight, that is your first fix.
2. Seller talk time. Most conferencing tools report it. Move it down five points before changing anything else.
3. Next step booked live. The share of discovery calls ending with a dated meeting in the calendar. The best leading indicator of whether a pipeline is real.
4. Discovery to second meeting conversion. If this is very high, your bar is too low and you are advancing deals you should have ended. A rising disqualification rate usually means the process is working.

Start here this week. Pull the recordings of your last five discovery calls, count the questions you asked, estimate your talk time, and note whether you left with a number, a name and a date. Most sellers find they asked four to six questions and talked for two thirds of the call. Fix the question count first, because it drags talk time down on its own. Then run the twelve questions on your next call exactly as written, before adapting them. More on pipeline that holds under pressure sits in the Sales Growth library.