Back to skills
SKILL.md
Gtm Win Loss
ASecurityBuild battlecards, run loss interviews, code reasons. Use when losing to incumbent, discounting blind, or message untested. For positioning, see product-marketing.
- 5 stars
- 0 votes
- 0 copies
- 1 view
- Added September 19, 2026
Works with
Security analysis
100/100Pro scans all 2 files and shows the line behind each finding
npx -y skills add DoddleHQ/DoddleOS --skill gtm-win-loss --agent claude-codeAre you the author of Gtm Win Loss?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/doddlehq-gtm-win-loss)---
name: gtm-win-loss
id: doddle.gtm.win-loss
version: 1.0.0
blueprint: ./blueprint.yaml
description: Build battlecards, run loss interviews, code reasons. Use when losing to incumbent, discounting blind, or message untested. For positioning, see product-marketing.
---
# GTM Win-Loss
Traps + proof + loss loop. Turns losses into message and product fixes.
## Language & Quality Standards
**CRITICAL**: Respond in same language user uses.
**Standards**: Token efficiency, sacrifice grammar for concision, list unresolved questions at end.
---
## When to Use This Skill
Apply when:
- Loss rate rising vs named competitor
- Reps discounting without guardrails
- No battlecard or stale (>6 mo)
- Product blind to loss reasons
## Initial Assessment
1. **Field**
- Top 3 competitors faced? Loss reasons hypothesized?
2. **Data**
- CRM loss notes usable? Interview access to buyers?
---
## Inputs Schema
| Input | Type | Required | Description |
|-------|------|----------|-------------|
| competitor | string | yes | Rival name; ask if missing |
| product_offer | string | no | Own offer + proof; infer if missing |
| locale | string | no | Market; default en-US |
---
## Outputs Schema
| Output | Type | Description |
|--------|------|-------------|
| report | markdown | Battlecard, traps, proof, interview script |
| action_list | json | Fixes with owner + effort |
---
## Core Framework
### Step 1: Map
Competitor claims → traps → own proof.
### Step 2: Card
One-page battlecard: landmines, questions, kill shots.
### Step 3: Loop
Interview → code → feed message + product.
---
## Detailed Guidance
### Battlecard
**Checklist:**
- [ ] Rival positioning + pricing (as known)
- [ ] 3 traps + discovery questions exposing them
- [ ] Proof: logos, numbers, migration stories
- [ ] When to walk away (bad-fit)
### Interviews
**Checklist:**
- [ ] 15-min script, third-party tone
- [ ] Code: price, features, trust, timing, incumbent
- [ ] 5 interviews per quarter minimum
### Feedback
**Checklist:**
- [ ] Message fixes → copywriter
- [ ] Product gaps → roadmap note
- [ ] Card refresh date stamped
---
## Common Mistakes
### Strategy Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| Trash-talk card | Loses trust | Facts + questions |
| No walk-away | Wastes cycles | Disqualify fast |
### Execution Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| 10-page card | Unused | One page |
| AE interviews own loss | Biased | Neutral interviewer |
### Analysis Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| Anecdote as trend | Misleads | Code 5+ first |
---
## Metrics to Track
| Metric | Definition | Target | Tool |
|--------|------------|--------|------|
| Win rate vs named | Wins / (wins+losses) | Improving | CRM |
| Discount rate | Avg discount | Guardrailed | CRM |
| Card usage | Reps opening | >70% | Enablement |
| Interview cadence | Per quarter | ≥5 | Manual |
---
## Decision Tree
No card → build for top rival first. Discount chaos → guardrails before cards. Low volume → interview every loss.
---
## Quick Assessment Checklist
1. [ ] Rival + context?
2. [ ] Loss notes access?
3. [ ] Buyer interview OK?
4. [ ] Proof assets?
5. [ ] Card owner?
---
## Expected Output Format
### Battlecard
[Traps, questions, proof]
### Interview Kit
[Script, codes]
### Fixes
[Message + product]
---
## Common Failure Modes
| Failure | Symptom | Fix |
|---------|---------|-----|
| Stale card | Wrong pricing | Expiry date |
| No adoption | PDF graveyard | 5-min drill |
---
## MCP Tool Integration
| Tool ID | When to Use | Data to Pull | Required |
|---------|-------------|--------------|----------|
| doddle.tool.v1.hubspot.deals | Loss coding | Stages, reasons | no |
| doddle.tool.v1.semrush.domainOverview | Rival sizing | Traffic, overlap | no |
Fallback: if tool unavailable, state data as NOT AVAILABLE per `data-reliability-rules.md`. Never fabricate.
---
## Agent Collaboration
| Agent | When to Collaborate | What They Provide |
|-------|--------------------|--------------------|
| researcher | Rival intel | Claims, pricing |
| sales-enabler | Card production | One-pager |
| copywriter | Trap phrasing | Questions |
---
## Related Skills
- **gtm-strategy**: Plan context
- **gtm-sales-playbook**: Guardrails
- **product-marketing**: Message depth
---
## Questions to Ask
1. Which competitor?
2. Loss notes available?
3. Interview access?
Files in this skill
- SKILL.md
- blueprint.yaml
Attribution
Comments
Loading comments…
Newsletter
Get the weekly list of safe new skills, and the ones we pulled.