Commercial Forecaster
Builds a defensible quarterly bookings forecast by decomposing pipeline into commit, best-case, and pipe-only tiers, projecting cohort-level NRR/GRR to surface leaky cohorts, and scoring per-stage funnel confidence. Every output includes an explicit assumption block naming conversion rates, data windows, and weighting choices, so the CRO can present numbers the board can trust.
✨ What it does
- Generates a 3-tier bookings forecast (commit, best-case, pipe-only) with disclosed conversion rates, data windows, and 7
- Projects per-cohort NRR/GRR over the horizon and flags leaky cohorts that consolidated numbers would hide.
- Scores each funnel stage with coefficient-of-variation confidence bands (HIGH/MEDIUM/LOW/VERY LOW) and treatment recomme
- Enforces a non-optional assumption block on every output; the CLI refuses to omit it.
- Includes a forcing-question library for grill-style validation of forecast assumptions.
- Provides pipeline-coverage ratio checks to flag structurally unsupported commits.
🎯 When to use it
- Preparing a quarterly bookings forecast or board deck where the CRO needs to present commit, best-case, and pipe-only numbers.
- Projecting ARR or NRR over the next 4-8 quarters and suspecting a consolidated NRR is hiding a leaky recent cohort.
- Assessing which pipeline stages are reliable enough to base forecasts on, especially when pipeline coverage is shrinking.
- Being asked for a single forecast number but needing a structured answer that surfaces assumptions and avoids theatre.
🚀 How to use
Trigger: /cs:grill-commercial
Trigger by asking for a quarterly bookings forecast, ARR projection, or pipeline forecast, especially when the CRO needs to present to the board. Fill the intake template with pipeline, cohort retention, and historical conversion data, then run the three scripts in sequence. Example prompts:
Build a Q3 bookings forecast with commit, best-case, and pipe-only numbers.
Project ARR for the next 6 quarters and show me which cohorts are leaking.
Score my funnel stages for confidence and tell me which are reliable for forecasting.
📄 Output: A forecast deck or report containing the 3-tier bookings number, cohort heatmap with leaky-cohort callout, funnel confidence scores, and an explicit assumption block.
📦 Add this skill to Claude Code
# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills
# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/commercial/skills/commercial-forecaster .claude/skills/commercial-forecasterSkill source: commercial/skills/commercial-forecaster/SKILL.md
⚠️ Good to know
Requires historical stage conversion data (last 4Q and 12Q) and cohort retention data; not for financial close, strategic planning, or pricing decisions.
❓ FAQ
What does the skill output?
It outputs three forecast numbers (commit, best-case, pipe-only) plus an assumption block, a cohort-level NRR/GRR projection with leaky-cohort callout, and per-stage funnel confidence scores.
Why does it use a 70/30 weighting between last-4Q and last-12Q conversion?
The blend captures recent regime change without overfitting to a single bad quarter; last-12Q alone hides slowdowns, while last-4Q alone overfits one quarter.
Can I remove the assumption block from the output?
No, the skill's hard rule is that a forecast without a disclosed assumption block is theatre; the CLI refuses to omit it.
🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.