sidchaudhary/gtm-skills / skills/gtm-engineer/call-notes
call-notes
UnverifiedReads a raw sales call transcript and extracts what the follow-up needs: deal signals, objections raised, pain points the prospect actually confirmed rather than ones the rep suggested, a stakeholder map, and one specific recommended next action. Use after a discovery call, before writing the follow-up email or updating the CRM record. Boundary: extracts deal facts for the follow-up and the CRM, while `call-preparation` grades the rep's own performance on that same call.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.3 KB |
| references/agent-security.md | markdown | 7.2 KB |
| references/coaching-metrics.md | markdown | 20.8 KB |
| references/house-rules.md | markdown | 13.3 KB |
references/coaching-metrics.md
View on GitHub# Sales Conversation Analysis
> **Provenance.** The numbers in this file are **pack benchmarks**, not the reader's own data, and
> most carry no named study behind them. When you state one in output, label it inline as a pack
> benchmark so nobody mistakes it for a figure derived from their business. Where a number is doing
> real work in a decision and you cannot name a source for it, write `[NEED: source]` instead of the
> number. House rule 4b covers this and binds every skill that reads this file.
Reference for analyzing sales conversations: talk ratio, patience, monologue detection, interactivity, question rate, BANT and MEDDIC qualification frameworks, objection detection, sentiment scoring, and coaching output structure.
---
## Talk Ratio Benchmarks
The percentage of time the sales rep talks versus listens during a conversation.
| Talk Ratio (Rep) | Rating | Description |
|-------------------|--------|-------------|
| < 30% | Too passive | Rep is not guiding the conversation or sharing enough value |
| 30-40% | Good for discovery | Appropriate for early discovery calls, mostly listening |
| 40-60% | Ideal range | Balanced conversation, the golden zone |
| 60-70% | Leaning monologue | Rep is talking too much; risk of losing prospect |
| > 70% | Monologue | Rep is dominating; prospect is disengaged |
### Talk Ratio by Call Type
| Call Type | Ideal Rep Talk % | Rationale |
|-----------|-----------------|-----------|
| Discovery | 30-40% | Listen to understand pain, let prospect talk |
| Demo | 50-60% | Presenting, but with interactive Q&A |
| Negotiation | 40-50% | Balanced, presenting offers, hearing concerns |
| Follow-up | 35-45% | Checking understanding, addressing questions |
| Closing | 40-55% | Summarizing, handling last objections, asking for the deal |
### Directional Patterns (No Single Authoritative Benchmark)
Conversation intelligence vendors publish talk-ratio and monologue research, but the exact percentages and thresholds vary by vendor, methodology, and year, and none of it is independently reproducible across sources. Treat the following as directional, not guaranteed:
| Pattern | Direction | Caveat |
|---------|-----------|--------|
| Top performer talk ratio | Tends to run lower than average reps, roughly in the 40s% | Varies by call type and by which study you read; no single number is authoritative |
| Questions per call | Top performers generally ask more, and more open-ended, questions than average reps | Exact counts vary widely by industry and deal size |
| Longest monologue | Shorter uninterrupted talk stretches correlate with better outcomes across most published research | Specific second thresholds differ by source and are not consistently reproducible |
| Multi-threading impact | Deals with multiple engaged stakeholders win at meaningfully higher rates than single-threaded deals | Magnitude (whether that is roughly 2x or higher) varies significantly by deal size and source |
| Collaborative language ("we/us") | Directionally associated with better outcomes in most vendor research | No consistent multiplier across studies; treat as a coaching signal, not a scored metric |
| Pricing timing | Introducing pricing after value is established tends to correlate with higher win rates | Magnitude not consistently reproducible across sources; use as qualitative guidance |
---
## Patience Score
Measures how long the rep waits in silence before interrupting or filling the gap.
| Silence Duration | Rating | Interpretation |
|-----------------|--------|---------------|
| > 5 seconds | Excellent | Comfortable with silence; prospect feels space to think |
| 3-5 seconds | Good | Adequate patience; allows prospect to complete thoughts |
| 2-3 seconds | Average | Acceptable but could improve |
| 1-2 seconds | Below average | Jumping in too quickly; prospect may feel rushed |
| < 1 second | Poor | Interrupting or overlapping; prospect will disengage |
### Measurement
- Track the gap between the prospect finishing a statement and the rep starting their next statement
- Average across all conversation turns
- Identify instances of overlap (negative silence = interruption)
- Flag interruptions separately: count of times rep spoke before prospect finished
---
## Longest Monologue
The longest uninterrupted stretch of the rep talking without asking a question or receiving input.
| Duration | Rating | Impact |
|----------|--------|--------|
| < 30 seconds | Excellent | Highly conversational, engaged prospect |
| 30-60 seconds | Good | Appropriate for demos and presentations |
| 60-90 seconds | Caution | Getting long, prospect attention may wander |
| > 90 seconds | Problem | Losing prospect attention; needs to break up with a question |
| > 120 seconds | Critical | Almost certainly lost the prospect; monologue mode |
### Remediation
When longest monologue > 90 seconds, coaching recommendation:
- Insert a check-in question every 60 seconds: "Does that make sense so far?" or "How does that compare to what you're doing today?"
- Break presentations into segments with interactive transitions
- Use the "teach-ask" pattern: share an insight, then ask how it relates to their situation
---
## Interactivity Score (0-100)
A composite metric measuring how conversational and engaged the call is.
### Components
| Component | Weight | Measurement |
|-----------|--------|-------------|
| Question rate | 30% | Questions asked per 10 minutes |
| Turn-taking frequency | 25% | Number of speaker switches per minute |
| Balanced talk ratio | 20% | Proximity to ideal 50/50 split |
| Prospect engagement | 15% | Length and depth of prospect responses |
| Silence comfort | 10% | Average pause duration before response |
### Score Ranges
| Score | Label | Description |
|-------|-------|-------------|
| 80-100 | Highly Interactive | Dynamic, collaborative conversation with balanced exchange |
| 60-79 | Interactive | Good back-and-forth with occasional long stretches |
| 40-59 | Moderately Interactive | Some dialogue but too much one-way talking |
| 20-39 | Low Interactivity | Dominated by one speaker; feels like a presentation |
| 0-19 | Monologue | One-way communication; not a conversation |
---
## Question Rate
The number of questions the rep asks per 10-minute window.
| Rate | Rating | Description |
|------|--------|-------------|
| 5+ per 10 min | Over-questioning | Feels like an interrogation; balance with insights |
| 3-5 per 10 min | Ideal | Genuine curiosity driving productive dialogue |
| 2-3 per 10 min | Adequate | Room for improvement; should probe deeper |
| 1-2 per 10 min | Low | Presenting more than discovering |
| < 1 per 10 min | Critical | Not asking questions; cannot qualify or understand needs |
### Question Quality Taxonomy
| Question Type | Quality | Example |
|--------------|---------|---------|
| Open-ended | High | "What's driving the need to solve this now?" |
| Probing | High | "Can you tell me more about how that impacts your team?" |
| Clarifying | Medium-High | "When you say 'too slow,' what specifically do you mean?" |
| Confirming | Medium | "So the main issue is X, is that right?" |
| Closed (yes/no) | Low-Medium | "Are you using [tool] today?" |
| Leading | Low | "You'd agree that faster is better, right?" |
| Loaded | Avoid | "Why haven't you fixed this already?" |
---
## BANT Framework
### Budget
Does the prospect have budget allocated for this purchase?
| Score | Criteria |
|-------|----------|
| 25 | Budget explicitly confirmed and allocated for this fiscal period |
| 20 | Budget exists but not yet allocated; needs internal approval |
| 15 | Budget identified but competing priorities; requires reallocation |
| 10 | No budget currently; exploring for future planning |
| 5 | Budget constrained; not a priority for allocation |
| 0 | No discussion of budget; unknown |
### Authority
Is the contact the decision maker or can they influence the decision?
| Score | Criteria |
|-------|----------|
| 25 | Contact is the final decision maker with signing authority |
| 20 | Contact is a strong champion who directly influences the decision maker |
| 15 | Contact is part of the buying committee but not the primary authority |
| 10 | Contact can recommend but has limited influence on the decision |
| 5 | Contact is an end user with no purchasing influence |
| 0 | Decision-making authority is unknown |
### Need
Is the pain real, acknowledged, and urgent?
| Score | Criteria |
|-------|----------|
| 25 | Explicit, urgent pain with quantified business impact |
| 20 | Clear pain acknowledged by multiple stakeholders |
| 15 | Pain identified but not yet quantified or broadly acknowledged |
| 10 | Vague dissatisfaction with current state; exploring options |
| 5 | Latent need; prospect not actively looking to solve |
| 0 | No need identified or discussed |
### Timeline
When does the prospect need a solution in place?
| Score | Criteria |
|-------|----------|
| 25 | Active evaluation with decision deadline within 30 days |
| 20 | Defined timeline: decision within 60-90 days |
| 15 | General timeline: "this quarter" or "this half" |
| 10 | Vague timeline: "sometime this year" |
| 5 | No timeline pressure; "looking for the future" |
| 0 | Timeline not discussed |
### BANT Total Score
| Total (0-100) | Qualification |
|---------------|--------------|
| 80-100 | Highly Qualified, pursue aggressively |
| 60-79 | Qualified, continue working, address gaps |
| 40-59 | Partially Qualified, needs development in weak areas |
| 20-39 | Weakly Qualified, nurture, check back later |
| 0-19 | Not Qualified, disqualify or long-term nurture |
---
## MEDDIC Framework
### Metrics
What is the quantified business impact of solving this problem?
| Score | Criteria |
|-------|----------|
| 17 | Clear, quantified metrics agreed upon: "save $500K/year" or "reduce churn by 3%" |
| 13 | Metrics identified but not yet agreed or verified by prospect |
| 9 | General impact discussed: "improve efficiency" or "grow revenue" |
| 5 | Impact mentioned vaguely; no quantification attempted |
| 0 | Business impact not discussed |
### Economic Buyer
Who signs the check? Have you engaged them?
| Score | Criteria |
|-------|----------|
| 17 | Economic buyer identified, engaged, and supportive |
| 13 | Economic buyer identified and aware; not yet directly engaged |
| 9 | Economic buyer identified but access is blocked or delayed |
| 5 | Economic buyer role is known but specific person is not identified |
| 0 | Economic buyer unknown |
### Decision Criteria
What are the technical and business requirements for the solution?
| Score | Criteria |
|-------|----------|
| 17 | Formal evaluation criteria documented and your solution maps well to all |
| 13 | Criteria known and you meet most; actively addressing gaps |
| 9 | Some criteria known; unclear if your solution is the best fit |
| 5 | Vague criteria; "we'll know it when we see it" |
| 0 | Decision criteria not discussed |
### Decision Process
What are the stages, timeline, and stakeholders involved in making the decision?
| Score | Criteria |
|-------|----------|
| 17 | Full process mapped: stages, timeline, stakeholders, approval chain |
| 13 | Process partially understood; key stages and timeline known |
| 9 | General understanding of process; some stages unclear |
| 5 | Vague: "we'll discuss internally and get back to you" |
| 0 | Decision process not discussed |
### Identify Pain
What specific pain points has the prospect articulated?
| Score | Criteria |
|-------|----------|
| 17 | Multiple specific pains identified with business impact quantified for each |
| 13 | Primary pain clearly articulated with some impact understanding |
| 9 | Pain acknowledged but not deeply explored or quantified |
| 5 | General dissatisfaction expressed; no specific pain points |
| 0 | No pain discussed; prospect seems satisfied with status quo |
### Champion
Is there an internal advocate with power, influence, and motivation to push the deal?
| Score | Criteria |
|-------|----------|
| 17 | Strong champion identified: has power, vested interest, and actively sells internally |
| 13 | Champion identified with influence but limited organizational power |
| 9 | Potential champion, supportive but untested in advocacy role |
| 5 | Friendly contact but no indication of willingness to champion |
| 0 | No champion identified |
### MEDDIC Total Score
| Total (0-102) | Qualification |
|---------------|--------------|
| 80-102 | Strong Deal, high probability of close |
| 60-79 | Good Deal, address remaining gaps proactively |
| 40-59 | Developing Deal, significant work needed on weak elements |
| 20-39 | At Risk, major gaps that threaten the deal |
| 0-19 | Weak Deal, consider deprioritizing or disqualifying |
---
## Objection Detection
### Objection Categories
| Category | Keywords/Phrases | Example |
|----------|-----------------|---------|
| **Price** | "too expensive", "budget", "cost", "pricing", "can't afford", "ROI", "cheaper" | "Your pricing is higher than what we're paying now" |
| **Timing** | "not now", "next quarter", "not a priority", "busy", "revisit later", "bad timing" | "We're focused on other things right now" |
| **Competition** | "already using", "competitor", "evaluated", "compared", "alternative", "other vendors" | "We're already working with [competitor]" |
| **Status Quo** | "working fine", "don't need", "not broken", "current solution", "happy with" | "What we have works well enough" |
| **Authority** | "need to check", "not my decision", "talk to my boss", "committee", "approval" | "I'd need to get buy-in from my VP" |
| **Trust/Risk** | "concerns about", "security", "implementation", "disruption", "change management" | "We're worried about the migration effort" |
| **Feature Gap** | "does it do", "missing", "need", "require", "can't do" | "Does it integrate with our [specific tool]?" |
### Objection Scoring
| Score | Interpretation |
|-------|---------------|
| 0 objections detected | Either too early in conversation or prospect is very receptive |
| 1-2 objections | Normal, healthy sign of engagement |
| 3-4 objections | Moderate concern, ensure each is addressed |
| 5+ objections | High concern, prospect may not be a fit or rep is losing control |
---
## Sentiment Scoring
### Breakdown
| Sentiment | Definition | Indicators |
|-----------|-----------|------------|
| **Positive** | Prospect is engaged, interested, enthusiastic | Asking follow-up questions, expressing interest ("that's interesting"), affirming ("yes, exactly"), discussing next steps, using future tense ("when we implement") |
| **Neutral** | Prospect is listening but not committed | Short responses, factual questions, no emotional language, neither enthusiastic nor resistant |
| **Negative** | Prospect is resistant, skeptical, or disengaged | Objections, pushback, skeptical questions ("but what about..."), short/dismissive responses, disengagement cues ("I need to go"), challenges ("prove it") |
### Sentiment Trajectory
| Pattern | Interpretation | Action |
|---------|---------------|--------|
| Positive → Positive | Strong engagement throughout | Move to next stage confidently |
| Neutral → Positive | Warming up; message is landing | Continue current approach, accelerate CTA |
| Positive → Negative | Something went wrong mid-call | Identify the trigger, address the concern directly |
| Negative → Positive | Objection successfully handled | Reinforce the resolution, build on positive momentum |
| Negative → Negative | Not landing; prospect is not a fit or approach is wrong | Reconsider qualification, try a different angle, or gracefully exit |
---
## Coaching Output Structure
Every conversation analysis should produce a structured coaching report.
### 1. Strengths
What the rep did well, reinforce these behaviors.
```
Strengths:
- Strong opening: established credibility in first 30 seconds with a relevant trigger event
- Excellent discovery questions at 4:20 and 7:15, open-ended and pain-probing
- Good patience score (4.2s average), let the prospect think
- Handled pricing objection at 18:30 with quantified ROI response
```
### 2. Weaknesses
Areas for improvement, specific, actionable, not personal.
```
Weaknesses:
- Longest monologue of 2:15 at 12:00-14:15, lost prospect attention (their responses shortened after)
- Talk ratio skewed to 65%, too much presenting, not enough dialogue
- Did not ask about timeline or budget, BANT incomplete
- Missed opportunity to multi-thread when prospect mentioned "my team would need to see this"
```
### 3. Specific Rewrites (with Timestamps)
Concrete, word-for-word alternatives for specific moments.
```
Rewrites:
[12:30] Instead of: "Let me walk you through all of our features..."
Try: "Which of these areas is most relevant to your situation? I'll focus there."
[18:45] Instead of: "Our pricing is very competitive."
Try: "Teams your size typically see a 3x ROI within 6 months, let me show you the math."
[22:10] Instead of: "Let me know if you have any questions."
Try: "Based on what we discussed, what would need to be true for you to move forward?"
```
### 4. Recommended Drills
Targeted practice exercises to build specific skills.
| Drill | Description | Duration | Frequency |
|-------|-------------|----------|-----------|
| **60-second teach-ask** | Present a value point in 60 seconds, then ask a question | 15 min | 2x/week |
| **Question ladder** | Practice asking 3 escalating questions on any topic | 10 min | Daily |
| **Objection handling** | Role-play top 3 objections with a partner | 20 min | Weekly |
| **Silence tolerance** | Practice 5-second pauses after asking a question | 10 min | Daily |
| **BANT check** | After every mock call, score BANT from memory | 5 min | After every call |
| **3x3 research** | Practice finding 3 personalized insights in 3 minutes | 3 min | Before every call |
| **Closing pivot** | Practice transitioning from any topic to a next-step ask | 10 min | 2x/week |
---
## What the Conversation Data Actually Supports
These are the three metrics most call-coaching tools report, with what the large-sample data says and,
more importantly, how strong a signal each one really is. Coaching a weak signal hard is how a rep ends
up optimising a dashboard instead of a conversation.
### Talk ratio: a real effect, and a small one
The widely quoted target is around **43:57 rep-to-prospect**. The measured reality across large call
samples is closer to a **60/40** average, and, the part that usually gets left out, **the gap between
won and lost deals is about 5 percentage points**: roughly **57% rep talk time on closed-won** against
**62% on lost**.
Two consequences:
- **Talk ratio is a weak discriminator on its own.** A five-point spread means it separates won from
lost only faintly. A rep at 62% is not failing, and moving them to 57% is a marginal intervention.
Presenting it as the headline fix overstates it.
- **The tail is where it matters.** Above roughly **65%** rep talk time, win rates fall off more
clearly. So the useful coaching target is not "hit 43%", it is "do not exceed 65%", which is a much
easier and more honest ask.
Treat talk ratio as a **screen**, not a diagnosis: it flags calls worth listening to, and the reason the
call was lost is in what was said, not in the ratio.
### Discovery questions: fewer, not more
This one runs against the usual advice. In calls that closed, top reps asked around **15-16 questions**;
in lost calls the count was nearer **20**. Other analyses put the effective range at **11-14 questions
in a 30-minute discovery**.
More questions correlate with *worse* outcomes, and the mechanism is legible: a long question list gets
worked through like a checklist, which produces short answers and no follow-through. Fewer, open
questions that build on the previous answer produce longer answers and more usable information.
So **never coach "ask more discovery questions" from a low question count alone.** Ask whether the
questions were open, and whether each one followed from the last. A rep asking 20 closed questions has a
worse call than one asking 12 that build.
### Monologue length: the one with a clean threshold
Top reps keep continuous uninterrupted speech under about **76 seconds**. This is the most actionable of
the three because it is concrete, easy to observe, and easy to change.
And the inverse metric matters more than the direct one: the **longest uninterrupted prospect
monologue** correlates with engagement and deal progression. The goal is not primarily to shorten the
rep's monologue, it is to lengthen the prospect's. A rep who talks in short bursts but never lets the
prospect run is still not getting discovery.
### How to weight these when coaching
Rank by signal strength, not by ease of measurement:
1. **Longest prospect monologue**, did they ever really talk?
2. **Rep monologue over 76 seconds**, concrete, fixable
3. **Question quality**, open, and building on the last answer
4. **Talk ratio**, a screen, and only the >65% tail is a clear finding
5. **Question count**, nearly useless alone, and misleading if coached upward
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |