- Home
- Skills
- Research & Analysis
- Growth Review: Zero-Cost Revenue Playbook
More screenshots
Works with the AI Agents you already use
Growth Review: Zero-Cost Revenue Playbook
Find the one thing holding your revenue back, from your own Stripe, Paddle, Lemon Squeezy, Gumroad or Shopify exports, and get a 90-day playbook of moves that cost time, not money. 8 Python check scripts.
$19
Growth Review: Zero-Cost Revenue Playbook
Example session with this skill installed
We're an AI meeting-notes tool for small agencies, 19/49/129 USD a month, no ad spend. I want 48,000 USD MRR within a year and have about 15 hours a week. Attached: our Stripe payments export (all statuses), AI inference cost per customer for three months, and monthly funnel counts. Review our growth and write a playbook.
- Read your context and instructions
- Compiled the growth review
I ran check_exports.py, revenue_metrics.py, cohort_retention.py, funnel_check.py, unit_economics.py, pricing_check.py and revenue_levers.py on your exports.
The answer
MRR is 24,182 USD and growing 5.7% a month, but at today's acquisition and churn it levels off near 42,900 USD, below your 48,000 goal. The constraint is early retention: 65% of a year's customer losses happen by month 3, and only 33% of sign-ups connect a calendar. A quarter of churn (42 of 165 in six months) follows a failed card: fix that first, it is free.
Playbook (zero cash)
| # | Move | Evidence | Hours | Read on | Stop if |
|---|---|---|---|---|---|
| 1 | Payment retries, card-update and expiry emails | 25% of churn follows a failed payment | 2 once | 1 Jan | complaints about the emails |
| 2 | A finished sample note before calendar setup; day 1, 2 and 5 emails to non-connectors | 33% connect a calendar | 5 a week | 11 Jan | calendar rate not above 36% |
| 3 | Fair-use limit on Team, a plan for heavy users | 14 customers cost more than they pay | 3 a week | 2 Feb | more than 3 of 14 cancel |
| 4 | Grandfathered Pro (29 USD) to 49 with 60 days' notice | 26 customers, about 500 USD MRR | 2 once | 31 Mar | more than 5 of 20 cancel |
The goal
With moves 1 and 2 working (churn down 30%, half of failed cards recovered), month-12 MRR is about 40,000 USD. 48,000 also needs about 30% more new customers a month; that combination crosses it in month 13. A/B testing the calendar step would take about 31 weeks at your traffic, so we measure before and after against your six-month history.
Connects securely to your tools. The creator never sees your data.
About this skill
Most growth advice is a list of tactics. This skill starts with your numbers: it reads the billing exports you already have, finds the single constraint on your revenue, and turns it into 3 to 5 moves that cost hours, not money, each with a metric and a date to judge it. It works in Claude Code, Codex and any agent that reads SKILL.md.
What it does
- Measures where revenue comes from and where it leaks. MRR and its movements (new, expansion, reactivation, contraction, churn), churn that follows a failed card, net and gross revenue retention, the quick ratio, and concentration. For shops and one-off products: repeat purchase, days to second order and returning revenue.
- Finds the ceiling. The level where MRR stops growing at today's acquisition and churn. If your goal is above it, the skill says so plainly and models which combination of levers gets there, and by which month.
- Shows where customers leave. Cohort retention by start month, the funnel step that loses the most people who already signed up, and whether churn is front-loaded (an onboarding problem) or steady (a value problem).
- Checks the money per customer. Margin per customer and per plan, including AI inference costs: the heavy users who cost more than they pay, and the free or churned accounts still using the product.
- Finds revenue you already earned. Failed payments nobody retried, customers still on old prices, discounts nobody ended, with the break-even for each price move.
- Writes the playbook. Each move has the evidence, cash cost (zero), hours a week, the metric today, a read date and a stop rule. It drafts the first one (dunning emails, a price-change notice, an onboarding sequence) for you to approve.
8 scripts, standard library only, no network
check_exports.py (what your exports contain), revenue_metrics.py, cohort_retention.py, funnel_check.py, unit_economics.py, pricing_check.py, revenue_levers.py (which lever moves revenue most, with scenarios) and experiment_plan.py (whether an A/B test can work at your traffic). Stripe payments, subscriptions and invoices, Paddle, Lemon Squeezy, Gumroad, Shopify and Recurly exports are detected automatically; any CSV with a customer, a date and an amount works.
Honest benchmarks
Every benchmark was checked at the organisation that collected the data (November 2026), with sample, date and an evidence grade, plus a list of popular figures that turned out to be unsourced or misquoted. Your own history comes first; benchmarks are context.
Private by design
Customer ids and emails are hashed on load and never printed. The scripts make no network requests and read no environment variables.
What it will not do
Invent benchmarks or a growth score, promise revenue, recommend ad spend you said you do not want, or email customers, change prices or edit billing settings without your go-ahead.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Research & Analysis.
Works from your exports, so its numbers are only as complete as they are: failed payments must be in the export to measure involuntary churn, and 13 months are needed for 12-month retention. MRR from payment exports is an approximation; reconcile it with your billing platform's MRR. The levers model is arithmetic on today's rates, not a forecast. Benchmarks were checked in November 2026; re-check anything older than six months. Not legal, tax or accounting advice.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
Trust & safety
Security scan
Security scan passed 4 days ago
- Accounts
- Google Analytics MCP
- Software
Python 3.8+ (standard library only)
- Optional
- Per-customer AIorusage costs (CSV)Monthly funnel counts (CSV)
Bring a billing CSV export: Stripe, Paddle, Lemon Squeezy, Gumroad, Shopify, Recurly, or any CSV with a customer, a date and an amount.
Permissions
Shell: runs the bundled Python scripts on your exports. Read files: billing, cost and funnel exports you provide. Write files: the review and the scripts' report files, only where you ask. Network: none; the scripts make no requests. Environment variables: none are read. Customer ids and emails are hashed on load and never printed.