discovery_call_playbook v1.0.0

---

name: discovery_call_playbook description: Run B2B discovery calls end to end: pre-call research, SPIN question sequencing, SPICED note capture, MEDDPICC qualification for complex deals, a booked next step, same-day follow-up, and a scored call review. Use when preparing for, running, transcribing, or reviewing a discovery, qualification, or first sales call, or when a deal stalls after a promising first conversation.

---

Discovery Call Playbook

Purpose

A discovery call has one job: decide, with evidence, whether a real problem exists that is worth solving, what it costs, when it must be solved, and how the buyer will decide. It is not a demo and not a pitch with questions in front of it.

This skill covers the full arc: research before the call, a question framework matched to deal size, notes that capture the buyer's own words, a qualification call you can defend in a pipeline review, follow-up the same day, and a review checklist that shows what to fix on the next call.

Pick the framework by deal complexity, not by habit:

| Situation | Framework | Why | |---|---|---| | Any first call, buyer has not framed the problem yet | SPIN sequence | Moves the buyer from facts to stated need in their own words | | SaaS or recurring revenue, mid-size deal | SPICED capture | Adds Impact and Critical Event, the two fields stalled deals usually lack | | Complex deal, multiple stakeholders, long cycle | SPICED on the call, MEDDPICC behind it | MEDDPICC is a qualification checklist, not a script. Use it after the call to find gaps |

Full question bank: references/question-bank.md. Review rubric: references/call-review-checklist.md. Note template: assets/discovery-notes-template.md. Scorer: bin/discovery_score.py.

Workflow

1. Pre-call research (about 20 minutes, scaled to deal size)

Work in this order. Stop when you have enough to form hypotheses, not a dossier.

1. CRM and prior touchpoints first (about 3 minutes). Previous conversations, emails, content the prospect engaged with, and notes from colleagues outweigh anything you find externally. Walking in unaware of a prior conversation is the fastest way to lose trust.
2. Company, role, and recent news (about 7 minutes). What the company does, size and stage, the contact's actual responsibilities, and any trigger event: funding, hiring, expansion, leadership change, contract renewal, a deadline they have announced.
3. Logistics check (about 1 minute). Meeting link works, names and titles of everyone attending are right, time zone is right.
4. One specific objective (about 4 minutes). Write it as a testable sentence: "Confirm whether a dated event forces a decision this quarter" beats "have a good conversation."
5. Two or three hypotheses (about 5 minutes). For each likely problem, draft one Problem question and one Implication question in advance. Implication questions are hard to improvise well. Write them before the call.

Output of this step: a half-page brief with what you know, what you do not know, your objective, and your prepared questions. Never ask a Situation question you could have answered from public sources.

2. Open with an upfront contract (first 2 minutes)

State four things, then get a verbal yes:

• Time: how long the call will run.
• Purpose: you are here to understand their situation and judge fit, not to present.
• Agenda: their current setup, what is not working, what it affects, how a decision would happen.
• Outcome: at the end, you will agree together whether a next step makes sense, and book it on this call if it does.

Plain version: "Thanks for the time. We have 30 minutes. My goal is to understand how you handle this today and what is not working, so we can judge together if there is a fit. I will ask a lot of questions and pitch nothing. At the end, if it makes sense, we will book the next conversation before we hang up. Sound fair?"

3. Run the conversation with SPIN sequencing

Ask in this order. The sequence is the method: each type earns the next.

1. Situation (few, and only what research could not answer). Current process, tools, team, volume. Two or three at most. Every extra Situation question taxes the buyer's patience and returns nothing to them.
2. Problem (the core of the call). Where the current approach breaks, frustrates, or falls short. Stay on one problem until it is concrete: last occurrence, frequency, who is affected. Cover 3 to 4 problems at most in one call.
3. Implication (where urgency comes from). Consequences of each problem: cost, time, effect on other teams, effect on their own targets, what happens if it runs another two quarters. A problem the buyer has lived with for a year does not move until its consequences are said out loud, by them.
4. Need-Payoff (let the buyer state the value). Ask what changes if the problem goes away, and what that is worth. Value stated in the buyer's words persuades their colleagues later. Value stated in your words does not.

Working rules for the conversation:

• Spread questions through the whole call. Front-loading all questions in the first ten minutes turns the call into an interrogation.
• Aim for roughly 11 to 16 targeted questions total, not a long list worked through mechanically. Gong's analysis of 326,000 calls found won calls averaged about 15 to 16 seller questions and lost calls about 20, so volume is not the lever. Relevance is.
• Talk less than the buyer. In the same Gong dataset the average seller talked about 60 percent of the time, and closed-won calls sat lower than lost calls. Treat that as directional evidence for listening more, not as a quota to game.
• Follow the answer, not the script. Summarise in the buyer's words and check: "Did I get that right?" Then ask the next question their answer opened up.
• No demo, no slides, no pricing defence before Implication and Need-Payoff are done. If the buyer pushes for product early, answer briefly and return to discovery: "Happy to show you. So I show the right part, can I ask how that situation plays out today?"

4. Capture notes in SPICED fields, live

Fill assets/discovery-notes-template.md during the call, or from the recording within an hour while wording is still recoverable.

• Situation: current process, tools, team, scale. Facts only.
• Pain: the specific problem, in the buyer's words, quoted where possible.
• Impact: the number. Cost, hours, revenue at risk, or the metric they will use to justify a purchase internally. If they cannot give a number, record their estimate and mark it as an estimate.
• Critical Event: the dated trigger. A renewal, a launch, a compliance deadline, a board date. "This quarter" is not a critical event. A date with a consequence for missing it is.
• Decision: who is involved, what criteria they will judge on, what steps stand between interest and a signed agreement.

Keep facts and your inferences in separate lines. A note that mixes the two corrupts every later review. Mark anything you assumed with inference: so the next reader can challenge it.

5. Qualify honestly before you advance the deal

Right after the call, score the notes:

python3 ~/workspace/skills/discovery-call-playbook/bin/discovery_score.py notes.md

The scorer checks SPICED completeness, flags a missing Critical Event or unquantified Impact, and warns on talk share and question count when you record them in the notes header.

For a complex deal, also run the MEDDPICC gap check from references/call-review-checklist.md: Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion, plus Paper Process and Competition on enterprise deals. An empty field is a named risk, not a detail to pick up later. The two gaps that kill the most forecast deals are no direct access to the Economic Buyer and a helpful contact mistaken for a Champion. A Champion sells internally when you are not in the room and can state why you should win. A coach only gives you information.

Then make the call, in writing, in the notes:

• Advance: pain is concrete, Impact is quantified, a Critical Event or a credible path to one exists, and the next step involves the right people.
• Advance with a gap plan: fit is real but one field is empty. The next step exists only to fill that field, and it has a date.
• Disqualify: no felt pain, no consequence for doing nothing, no access to anyone who can decide, or the problem is one your product does not solve. Say so on the call or in the follow-up, plainly. A fast, honest no protects both sides' time.

6. Close the call properly (last 5 minutes)

1. Recap in the buyer's words: situation, problem, impact, date. Ask what you got wrong.
2. State your honest read on fit, including anything you cannot do.
3. Book the next step before hanging up: named people, purpose, date and time on both calendars. "I will send some times" is not a next step.
4. Agree what each side will do before that next step, and by when.

7. Follow up the same day

Send one email, the same business day, built from the notes:

• Subject: plain and specific, naming the problem or the date, not "Great call!"
• Three to five lines recapping their problem and its impact, in their words.
• The agreed next step with the date, time, and who is attending.
• One useful thing attached or linked only if it speaks directly to the problem they named. No generic deck.
• One open question if a SPICED field is still empty, framed as completing the picture before the next call.

Log the notes to the CRM the same day, with the qualification call and the gap list. A deal that lives in a notebook is not in the pipeline.

8. Review the call within 48 hours

Run references/call-review-checklist.md against the recording or transcript. Score each item pass or miss, name the single worst miss, and write one concrete change for the next call. Review one call a week minimum. Reviewing only lost deals teaches you less than reviewing a mix of won, lost, and stalled calls, because the same discovery gaps show up in all three.

Output Contract

For a prep request, deliver: the half-page brief (known facts with sources, open questions, one objective, prepared Problem and Implication questions, likely stakeholders).

For a live or past call, deliver: completed SPICED notes with buyer quotes, the scorer output, an advance / gap-plan / disqualify decision with reasons, the booked next step, and the follow-up email draft.

For a review request, deliver: the completed review checklist, the top gap, and one specific behaviour change for the next call, stated as an action ("prepare two Implication questions per hypothesis before the call"), not as a trait ("listen better").

Operating Rules

1. Every factual claim about the prospect's company in a brief needs a source you actually checked. Unverified claims get cut or labelled as a hypothesis to test on the call.
2. Never invent numbers for Impact. Record the buyer's figure, mark estimates as estimates, and leave the field empty rather than filling it with a plausible guess.
3. The buyer's words outrank your paraphrase. Quote pain and impact statements verbatim in notes wherever the recording allows it.
4. Do not pitch before the buyer has stated the problem and its consequences in their own words. If time runs out before that point, book a second discovery call instead of compressing discovery into a pitch.
5. No next step booked on the call means the call is not finished. Chase emails are not a closing plan.
6. Disqualifying a bad fit is a success outcome of discovery, not a failure. Report it as one, with the evidence.
7. Respect recording consent and local law. State that the call is recorded where required, and never capture notes you would not show the buyer if asked.