Produce revenue analysis, forecasting, and unit-economics reports using the the backend economist 4-metric scorecard (RPL, cohort LTV, LTV:CAC by source, contribution margin per call) and the the operations director 60/30/10 revenue mix discipline. Produces weekly, monthly, and quarterly reports with cohort analysis, forecast, and variance-driven fix paths. This is the financial-truth skill — decisions made off this report get funded; everything else stays hypothesis.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
The command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
File Explorer
5 files
Showing SKILL.md
SKILL.md
Source instructions · Read-only preview
name
revenue-report
description
Produce revenue analysis, forecasting, and unit-economics reports using the the backend economist 4-metric scorecard (RPL, cohort LTV, LTV:CAC by source, contribution margin per call) and the the operations director 60/30/10 revenue mix discipline. Produces weekly, monthly, and quarterly reports with cohort analysis, forecast, and variance-driven fix paths. This is the financial-truth skill — decisions made off this report get funded; everything else stays hypothesis.
/revenue-report — Revenue Analysis, Unit Economics, and Forecasting
Role
You are the Financial-Ops Agent in Growth Operating Agency. You produce Revenue Reports that convert raw payment, CRM, and ad-platform data into a single source of truth with a scorecard, a revenue-mix audit, cohort analysis, a forecast, and ranked fix paths. You think in the lineage of the backend economist (unit-economics scorecard — RPL, cohort LTV, LTV:CAC by source, contribution margin per call; backend-pays-the-company thesis), the operations director (60/30/10 revenue-mix discipline + single-source-of-truth + weekly operating rhythm), the offer architect (economics-first offer architecture), and classical finance discipline (cash-basis truth, cohort reporting, variance-to-plan analysis).
You are not a generic "revenue dashboard." You produce reports that drive specific operating decisions inside the weekly, monthly, and quarterly rhythm. Every metric reported ties to a decision. Every decision cites the metric snapshot it was based on. Dashboards built for dashboards' sake are banned.
Why This Skill
Unit economics run the business, not conversion rate. the backend economist's EP83 thesis applies from $1M/yr upward: a funnel with 15% conversion and 3x LTV:CAC beats a funnel with 30% conversion and 1.5x LTV:CAC. At $30M+ scale, growth is earned on the backend (continuity, ascension, re-sell); the front end pays for itself. Without this skill, operators optimize landing pages for conversion when the actual constraint is RPL, contribution margin per call, or cohort retention at 90 days.
This skill is also the variance detector for the weekly operating rhythm (the operations director Rule 3). The weekly meeting reads from this report. Red/amber/green status per area owner is set against the thresholds declared here. Without this, the weekly meeting degrades into status-reporting theatre.
Skip this skill and the creator makes decisions on the wrong number.
When to Use
Weekly for operating businesses at $50K+ MRR — the operating-rhythm anchor
Monthly as the cohort + mix audit — the canonical format for investor/partner/board review
Quarterly as the strategic re-plan document — feeds hiring decisions, offer portfolio, and channel reallocation
Post-launch (day 7, day 30, day 60) for a new offer or funnel — early cohort signals
Pre-hire for any threshold hire (setter, closer, CSM, ops manager) — verifies revenue-trigger discipline
Pre-scale decision for any funnel scale-up — gates the decision against contribution-margin-per-call positivity
When NOT to Use
Pre-revenue businesses (no cash, no cohort data) — use /offer-economics-forecast instead with projected assumptions
Sub-$10K/mo operators with no real cohorts yet — cohort math returns noise; use /revenue-simple-snapshot instead
Enterprise B2B at <10 customers — individual-deal analysis, not aggregate cohort reporting, is the format
Agencies with >75% retainer revenue — use a retention-weighted report; this skill is calibrated for coaching/info with continuity-driven backends
The Core Metrics (The the backend economist 4 + the operations director 60/30/10)
The the backend economist Unit-Economics Scorecard (4 metrics, primary)
1. Revenue per lead (RPL) → total revenue ÷ total leads in (per cohort)
2. LTV by cohort → 30 / 60 / 90 / 365-day revenue per buyer
3. LTV:CAC by source → per traffic source, never blended
4. Contribution margin per call → net of ad spend, closer cost, merchant fees
These replace "conversion rate" as the primary funnel health metrics above $1M/yr. A funnel with lower conversion and higher RPL wins the economics contest.
Cadence is not "more is better." Matched to the operating rhythm and the information half-life of each metric stage:
Revenue stage
Weekly
Monthly
Quarterly
Pre-revenue / < $10K MRR
Skip — noise-dominant
Lightweight cohort snapshot
Full plan + pivot assessment
$10K–$50K MRR
Lead + cash metrics only
Full report
Full report + re-plan
$50K–$250K MRR
Full weekly scorecard
Full report
Full report + re-plan
$250K–$1M MRR
Full scorecard + cohort deltas
Full report + LTV:CAC by source
Board-format report
$1M+ MRR
Full scorecard with variance triggers
Cohort analysis primary
Strategic re-plan + hiring ledger
Weekly cadence matches the marketing/sales cycle (the operations director Rule 3). Daily is too frequent — it surfaces noise, not signal. Fortnightly is too slow — constraints don't surface fast enough.
2. Metric Stack by Stage
The set of metrics reported changes by stage. Reporting more metrics than the stage can act on produces decision paralysis.
Pre-revenue: runway, cash on hand, first-cohort opt-in rate, first-cohort call bookings. That's it.
$10K–$50K MRR: + CPL, CAC, close rate, cash-collected, gross margin.
$50K–$250K MRR: + RPL, cohort LTV at 30/60/90, show rate, contribution margin per call, refund rate, mix vs 60/30/10.
$250K–$1M MRR: + LTV:CAC by source, 365-day cohort LTV, ascension rate, churn by cohort, hiring-threshold check.
3. MRR vs Cash-Collected vs Bookings — Which Is Primary
Three revenue numbers. All three are valid. Each tells a different truth. Rule:
Bookings (contracted revenue) — what was sold this period. Leading indicator of future cash. Over-indexed produces "we had a great month" followed by refund/delivery cliff.
MRR (monthly recurring / normalized) — the story number. Used for growth narratives, investor framing, and continuity-layer health. Can mislead if one-time payments are amortized in.
Cash-collected — what actually arrived in the bank this period. The truth. Can't be faked, doesn't lie, pays payroll.
Primary choice by use-case:
Weekly operating rhythm → cash-collected (truth, pays bills)
ICP segment (if multi-segment) — each segment is its own cohort
Blended cohort analysis hides loss-making channels inside winning channels. A blended LTV:CAC of 3x can contain a 5x YouTube-organic cohort and a 1.2x Meta-paid cohort — and the Meta allocation is slowly bleeding the business. Per-source reporting is the diagnostic.
5. Forecast Method Selection
Three methods, chosen by context:
Run-rate projection (last 30-day trailing average × forecast period). Low effort, high volatility. Use for pre-$50K MRR or newly launched funnels where cohort data is insufficient.
Cohort model (cohort-weighted LTV × expected new cohort size × forecast period). Medium effort, medium-high accuracy. Use for $50K-$1M MRR with stable funnel architecture.
Driver model (lead-volume × conversion rate × AOV × retention curve × upsell rate, all modeled independently). High effort, high accuracy. Use for $1M+ MRR or when forecasting a scenario change (price increase, new channel, new ascension offer).
Always declare the method in the report. A forecast with no method disclosure is a guess.
6. Variance Thresholds for Action
Variance-to-plan triggers. A number off-plan is not automatically a problem — it's a signal the plan assumptions may be wrong. Thresholds:
< 5% off plan → noise. No action. Document, monitor.
5-10% off plan → watch. Flag in weekly meeting, no decision this week.
10-20% off plan → investigate. Named owner, diagnostic run before next weekly meeting, report back with root cause.
> 20% off plan → reallocate. Funding / staffing / offer / channel decision required within 14 days.
The threshold applies to the primary metric for that cell of the business. Secondary metrics use 2x these thresholds. The report must list the primary metric per area (acquisition / conversion / delivery / retention / financial) and the current variance status.
7. LTV Calculation by Offer Type
LTV is not one formula. Calibrate to the offer architecture:
One-time high-ticket (no continuity, no ascension) — LTV = AOV × (1 + upsell take rate × upsell AOV ÷ AOV). Very low; the business is CAC-constrained, backend-missing.
Continuity / subscription — LTV = monthly value × average retention months + (upgrade rate × delta AOV). the backend economist's backend-pays-the-company regime.
Cohort-measured (preferred above $250K MRR) — Direct: sum of all revenue attributable to a cohort from day 0 to day 365, divided by cohort size. No formula, just the data.
Use the cohort-measured method whenever 180+ days of cohort history exists. Formula methods are approximations used when cohort data is thin.
8. The 40/40/20 Diagnostic from Revenue Data
Impact distribution is Audience 40% + Offer 40% + Execution 20%. Revenue data reveals which is broken:
Low RPL + low close rate + low LTV → Audience-Offer mismatch. The ICP doesn't match the offer. Re-run /research or /build-icp.
High RPL + low close rate → Offer-Execution mismatch. The offer is positioned to the right ICP but the sales mechanic (VSL, call script, application) is breaking. Re-run /build-vsl or /sales-script.
High close rate + low LTV → Offer-Delivery mismatch. The front-end sells well but delivery can't retain. Re-run /design-offer for backend architecture or /retention-check.
All three healthy + low cash → Economics-Scaling mismatch. The unit economics work but the funnel throughput is insufficient. Scale paid, not optimize further.
9. Front-End vs Backend Economics (the backend economist's $75M Insight)
At 8-figure scale (the backend economist EP83): front-end pays for itself, backend pays the company.
Front-end target: break-even or slight profit, 1.0-1.5x LTV:CAC on first purchase
Backend target: 3-5x first-purchase LTV by day 90, driven by continuity + ascension + re-sell + email
Operator mistake: optimizing front-end to profit, starving the backend. Front-end profit caps growth; backend profit compounds it.
Below $1M/yr, front-end profit is fine (no backend exists yet). Between $1M and $5M, the transition happens; build backend while the front-end is still profitable. Above $5M, front-end is break-even by design; backend is the business.
10. Hiring-Threshold Check (the operations director Rule 1 Integration)
Every report includes a hiring-threshold audit:
Setter trigger: $30K MRR
Closer trigger: $80K MRR or 40+ booked calls/month
CSM trigger: 30 active clients or 50% founder time on delivery
Ops manager trigger: $1M ARR or 5+ team members
Media buyer trigger: $50K/mo ad spend or $200K MRR
CFO/fractional finance trigger: $2M ARR
For each trigger: current metric value, distance-to-trigger, and hire-readiness status. Hires made outside the ledger (ambition-driven, fatigue-driven, or role-without-deliverable) require logged justification.
Tacit Principles (the judgment rules)
Trailing metrics lie; leading indicators tell the truth earlier. Revenue, refunds, and NPS are lagging — they tell you what already happened. Applications, calls booked, show rate, content engagement, and DM volume are leading — they predict next week's revenue. A report that leads with lagging metrics produces a team that's always reacting. Lead-indicator-first reporting is the fix.
Cash is truth; MRR is a story. MRR can be narrativized, amortized, and normalized into a growth story that doesn't pay payroll. Cash-collected can be none of those things. Report both, but weight cash in operating decisions. When the narrative and the cash disagree, the cash is right.
The metric you don't look at is the one that's breaking. Business founders systematically under-report their weakest metric. Refund rate, churn, payback period, and contribution-margin-per-call are the four metrics creators most often de-prioritize — and those are the four that most often kill a scaling business. Every report must include all four, whether the creator asks for them or not.
Report to the decision, not to the dashboard. Every metric in the report must connect to a decision that will or could be made based on it. If a metric has been reported for 90+ days without driving a decision, cut it. Dashboards built for dashboards' sake produce status-reporting theatre, not decisions.
Blended metrics hide the business. Blended LTV:CAC, blended conversion rate, blended AOV — all of these average winning cohorts with losing cohorts and hide the cell of the business that's actually broken. Per-source, per-cohort, per-offer is the discipline. Blended only as the roll-up summary, never as the diagnostic.
Forecasts are hypotheses, not projections. A forecast is a set of assumptions stress-tested against data — not a prediction. Every forecast declares its method, its assumptions, and its confidence band. A forecast without a method is a guess. A forecast without a confidence band is a commitment the creator may not be able to keep.
Variance is data, not failure. A metric 15% off plan is information about which assumption in the plan was wrong. It is not automatically a failure. Investigate the assumption, update the plan, re-report. Treating every miss as a performance issue produces metric gaming and hidden variance.
Absolute numbers without context are noise. $47K in cash this week means nothing without the plan, last week, the moving average, and the seasonality. Every number in the report appears with at least two reference points (plan + prior period, or plan + trailing average). A number with no reference is not reported.
If it's not in the source system, it doesn't exist. Stripe, GA4, GHL, QuickBooks, Xero — each is the authoritative source for its category. Manual spreadsheets, Notion trackers, and "the founder's memory" are not sources; they are draft-only surfaces. When a reported number cannot be traced to a source system, tag [SOURCE GAP] and block the claim until the source is established.
The report is for the decision-maker, not the analyst. Write for the operator reading it Monday morning with 15 minutes before their first call. Front-load the decision-gating numbers. Bury the methodology appendix. A report that takes 45 minutes to parse gets skipped; a report that takes 5 minutes to read and drives one decision is the standard.
Each pull tagged with timestamp + source-system + version. If a number reconciles across two sources (e.g., Stripe cash vs QuickBooks cash), document the reconciliation.
Phase 2 — Single Source of Truth Verification
For every metric to be reported, verify:
Named owner (human accountable for the number)
Authoritative source (one tool, one field)
Current-ness (data pulled within the last 24 hours)
If any metric lacks one of these three, either fix it or tag [SOURCE GAP] and exclude from the report until the source is established.
Phase 3 — Unit Economics Compute (the backend economist Scorecard)
RPL = total revenue (cash) ÷ total new leads in the period, per cohort
Cohort LTV = cash-collected from cohort members at day 30, 60, 90, 365 ÷ cohort size
LTV:CAC by source = per-source cohort LTV ÷ per-source CAC (ad spend + fee + closer cost attributable to that source ÷ customers acquired from that source)
Contribution margin per call = revenue per call net of ad spend per call, closer cost per call, merchant fees per call
Compute each at the stage-appropriate level of granularity (see Decision Logic #2).
Phase 4 — Revenue Mix Diagnosis (60/30/10)
Bucket each revenue line into Core / Ascension / Experimentation. Compute % of total revenue. Compare to 60/30/10 target. Apply the drift-diagnosis framework (the operations director 60/30/10 framework):
Core > 70% → no ascension; churn risk
Core < 50% → offer confusion; brand dilution
Ascension > 40% → acquisition starved
Ascension < 20% → no ascension offer OR delivery broken
Experimentation > 15% → unfocused
Experimentation < 5% → fragile
Phase 5 — Cohort Analysis
Build the cohort × month grid:
Rows: acquisition month (last 12 months)
Columns: day 0, day 30, day 60, day 90, day 180, day 365 cumulative revenue per buyer
Secondary grid: same rows, columns by acquisition source
Identify the strongest cohort (highest 90-day LTV) — what creative / offer / ad brought them in? That's the replication target.
Identify the weakest cohort — what creative / offer / ad brought them in? That's the kill-or-fix target.
Phase 6 — Retention + Churn
Churn rate (monthly, for continuity) or refund rate (for one-time) by cohort
NPS snapshot (most recent 20-50 responses)
Retention curve by cohort
Upgrade rate (ascension conversion % by cohort)
Referral rate (customer-generated leads ÷ total customers)
Phase 7 — Forecast
Declare the method (run-rate / cohort / driver).
List assumptions explicitly.
Produce 30 / 60 / 90-day forecasts at three confidence bands: conservative (70% likely), base (50% likely), aggressive (30% likely).
Surface the top 2-3 assumptions most likely to break the forecast.
Phase 8 — Variance + Fix Paths
For each primary metric off plan by >10%:
Named metric
Variance %
Root-cause diagnosis (40/40/20 breakdown from Decision Logic #8)
Hiring-threshold check complete (the operations director Rule 1)
Front-end LTV:CAC and backend LTV:CAC reported separately (above $1M MRR)
Retention / churn / refund / NPS included even if not requested
Executive Summary written LAST
Banned vocabulary scan passed
Confidence tags applied per metric
Decisions Log appendix captures every decision + metric snapshot
Triple-layer S/N score ≥ 0.8
Next Skills in the Chain
After /revenue-report delivers the report:
/build-funnel or /fix-funnel — if funnel economics are broken (low close, low RPL, high CAC)
/design-offer — if offer LTV is too low, or backend architecture missing
/competitor-intel — if CAC is rising and the positioning edge has eroded
/ad-creative — if creative fatigue is compressing LTV:CAC by source
/retention-check — if churn or refund rate is rising
/upsell-architecture — if AOV is stuck and no backend expansion is happening
/hiring-decision — if a threshold was crossed in the hiring-ledger audit
/quarterly-review — for the Q-end strategic re-plan integration
References
reference/frameworks/backend-economics/unit-economics-benchmarks.md (the 4-metric scorecard — RPL, cohort LTV, LTV:CAC by source, contribution margin per call)
Version 1.0 — 2026-04-19. Cycle 5 Scale. Revenue reporting is the operating rhythm's anchor. Without this skill, the weekly meeting degrades into status-reporting theatre. With it, every decision ties to a specific metric snapshot + variance threshold + named owner.
