Agent skill
win-loss-decoder
Turns closed-won and closed-lost deals into ranked, evidence-backed reasons (not rep-selected dropdowns) and routes each pattern to the team that can fix it.
Filed under Calls, demos and discovery.
From varunk130/ai-gtm-skill-library · 37 skill entries · 6 · pushed 2026-09-19
What it does when it runs
Turns closed-won and closed-lost deals into ranked, evidence-backed reasons (not rep-selected dropdowns) and routes each pattern to the team that can fix it. Use when: win loss analysis, why did we lose, why do we win, closed lost review, loss reasons, churned deals, competitive losses, buyer interviews.
Automated analysis of the skill and the 0 files bundled beside it. A skill’s own description is written to be selected by an agent, so it describes the job and not the dependencies.
- Keys and connectors you must supply
- None found.
- Hosts it reaches
- No third-party host appears in the skill or its bundled files.
- Tool permissions it declares
- No
allowed-toolsin the frontmatter. It only issues instructions, so there is nothing to bound. - Actions present in the files
- None. Instructions only.
Install it
View source on GitHub ↗git clone --depth 1 --filter=blob:none --sparse https://github.com/varunk130/ai-gtm-skill-library.git /tmp/ai-gtm-skill-library git -C /tmp/ai-gtm-skill-library sparse-checkout set "gtm-skills/win-loss-decoder" mkdir -p ~/.claude/skills/win-loss-decoder cp -R "/tmp/ai-gtm-skill-library/gtm-skills/win-loss-decoder/." ~/.claude/skills/win-loss-decoder/
Picked up without a restart. A project skill of the same name is shadowed by your personal one. For one repository only, swap ~/.claude/skills for .claude/skills. Claude Code docs ↗
The folder is the same in every client that implements the format — 46 of them — so if yours is not above, only the destination changes.
The skill
Source on GitHub ↗Reproduced in full from varunk130/ai-gtm-skill-library/blob/158f12c03830dca454576eb1a78b1e4a34ae83e3/gtm-skills/win-loss-decoder/SKILL.md, which is licensed MIT (repository). 530 words, 8 headings.
Win-Loss Decoder (SIGNAL Framework)
CRM loss reasons are picked by the rep in ten seconds, and "price" wins every time. Win-Loss Decoder rebuilds the reason from what the buyer said and did, codes it against a fixed taxonomy, and ranks patterns by revenue so product, marketing, and sales each get the fixes that are theirs.
Core Principle
The buyer's words decide the code. A reason without a quote or artifact from the buyer is logged as "unknown", not guessed.
The SIGNAL Framework
| Letter | Step | The Question |
|---|---|---|
| S | Sample | Which deals, and is the sample balanced across wins, losses, segments, and reps? |
| I | Interview | What did the buyer say in a 20-minute conversation after the decision? |
| G | Group | Which taxonomy code does each reason map to? |
| N | Number | How much ARR sits behind each code, for wins and for losses? |
| A | Assign | Which team owns the fix for each code? |
| L | Loop | Did the fix change the rate of that code next quarter? |
Reason Taxonomy
| Code | Covers | Default owner |
|---|---|---|
| VALUE | Buyer could not see enough return or urgency | Marketing |
| FIT | Missing capability, integration, or compliance need | Product |
| TRUST | Proof, references, security, or vendor-risk concerns | Marketing / Security |
| PROCESS | Lost the sales process: slow, no champion, wrong buyer | Sales |
| PRICE | Price or packaging, after the above are ruled out | Pricing |
| STATUS QUO | Buyer chose to do nothing | Marketing / Sales |
Process
- Sample - 15-30 recent decisions, at least one-third wins; exclude deals under the ACV floor
- Collect - buyer interviews first, then call notes and emails; note the source of every reason
- Code - up to two codes per deal (primary, secondary); PRICE only if VALUE and FIT are ruled out
- Weight - sum ARR by code for wins and losses; compute
loss share - win shareper code - Rank - largest positive gap first; these are the reasons you lose more than you win on
- Route - one fix per top code, with owner and the metric that will show it worked
Output
Save to outputs/win-loss-decoder-[period]-[YYYY-MM-DD].md
| Artifact | Description |
|---|---|
| Coded deal log | Deal, outcome, codes, quoted evidence, source |
| Pattern table | Code, ARR won, ARR lost, gap |
| Fix plan | Top three codes with owner, action, and target metric |
Tips
- Interview wins too - you cannot tell a loss pattern from a market pattern without them
- Don't let reps interview their own deals - buyers are more candid with a neutral party
- Rerun quarterly - the Loop step is where the value compounds
Pairs With
- deal-xray - PROCESS losses usually show missing proof earlier in the deal
- competitive-battlecard - competitor-named losses update the battlecard
- position-lock - persistent VALUE losses signal a positioning problem
Other skills for the same job
Different authors, same problem. Matched on the words in the skill name, across every library in the catalogue except this one.
- deal-review-win-loss by louisblythe · 159
- win-loss-reason-extraction by louisblythe · 159
- win-loss-reasons by pmalliance · 69
- win-loss-analysis by matteotitta · 57
- win-loss-program by NEON-Rutger · 48
- win-loss-analysis by jbalbu01 · 15
- win-loss-brief by zime-ai · 15
- win-loss-report by octavehq · 11
Need help setting it up?
This page tells you what win-loss-decoder does and what it needs. Cheetah builds the agent setup it runs inside: data, CRM, sequencing and the guardrails.
Book a call →The directory stays free. There is nothing gated behind this.