# Build Monetization Loop > Turn a selected opportunity, expertise, product, service, or underperforming business into an evidence-backed end-to-end monetization loop covering customer problem validation, offer design, pricing and unit economics, acquisition, sales, fulfillment, payment, retention, referral, measurement, experimentation, and automation. Use for requests such as “设计商业变现闭环”, “把这个能力变成生意”, “从获客到复购做完整系统”, “诊断为什么… - **Type:** Skill - **Install:** `agentstack add skill-shangdizhiyan-vic-distill-industry-skills-build-monetization-loop` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [shangdizhiyan](https://agentstack.voostack.com/s/shangdizhiyan) - **Installs:** 0 - **Category:** [Finance & Payments](https://agentstack.voostack.com/c/finance-and-payments) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [shangdizhiyan](https://github.com/shangdizhiyan) - **Source:** https://github.com/shangdizhiyan/Vic-distill-industry-skills/tree/main/skills/build-monetization-loop ## Install ```sh agentstack add skill-shangdizhiyan-vic-distill-industry-skills-build-monetization-loop ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Build a Monetization Loop Turn a bounded opportunity into a measurable system that repeatedly creates, captures, delivers, and retains value. Treat revenue as a downstream result of verified customer behavior, sound economics, reliable delivery, and controlled learning—not as a promised outcome. ## Select the mode - `diagnose`: locate breaks in an existing funnel, economics, delivery, retention, or data loop. - `design`: create a complete monetization system from a selected opportunity or capability. - `validate`: test the riskiest assumptions with the smallest ethical experiments. - `launch`: prepare and execute an approved pilot with explicit owners, budgets, and stop rules. - `optimize`: improve a running loop using observed funnel, cohort, margin, and delivery data. - `scale`: automate or expand only after the relevant gates pass. - `audit`: inspect an existing monetization pack without rebuilding it. Choose the smallest mode that satisfies the request. Default to `design` only when the user clearly asks for a complete system. ## Establish the starting point Accept one of three valid inputs: 1. a dated opportunity and evidence pack from any credible research process; 2. an existing offer with customer, channel, sales, delivery, and financial data; or 3. a bounded capability plus a specific customer and problem hypothesis. If the user asks which industry or opportunity to enter, or the starting opportunity lacks industry-level evidence, stop the monetization workflow and request a bounded opportunity or a separate industry-research result. If another industry-research Skill is available, it may supply that input, but this Skill must not depend on one by name. Do not rebuild a whole industry inside this Skill. Freeze the business, customer, geography, offer type, operator role, resources, time horizon, currency, evidence cutoff, constraints, exclusions, and requested external actions. Separate known facts, user-provided claims, assumptions, and unknowns. ## Run the 12-stage loop Read [references/monetization-workflow.md](references/monetization-workflow.md) before `design`, `launch`, `optimize`, or `scale` work. Resume from `PIPELINE_STATE.md` when present. 1. **Scope the opportunity** — define the customer, problem, current alternative, value event, business boundary, and decision. 2. **Map assumptions and evidence** — record desirability, viability, feasibility, channel, trust, compliance, delivery, retention, and cash assumptions. 3. **Validate the customer problem** — collect behavioral evidence; distinguish interest, intent, commitment, payment, use, and repeat use. 4. **Design the offer ladder** — specify outcome, scope, mechanism, proof, exclusions, risk reversal, entry offer, core offer, and expansion path. 5. **Model price and economics** — calculate contribution margin, break-even volume, CAC ceiling, payback, capacity, cash timing, and scenario sensitivity. 6. **Design acquisition experiments** — choose channels by customer access and buying context; define testable messages, assets, budgets, and attribution. 7. **Build the sales system** — define qualification, discovery, proposal, objections, approval, contracting, payment, follow-up, loss reasons, and CRM states. 8. **Engineer fulfillment** — define onboarding, inputs, milestones, acceptance, quality, exceptions, handoff, support, capacity, and cost capture. 9. **Close the value loop** — measure activation, time-to-value, outcome, satisfaction, renewal, expansion, referral, and churn reasons. 10. **Instrument the system** — create one source of truth for lead, opportunity, order, customer, delivery, cash, experiment, and cohort events; reconcile booked revenue to collected cash. 11. **Run controlled iteration** — prioritize bottlenecks, change one major variable per experiment, record decisions, and update assumptions. 12. **Automate and scale** — automate stable, reversible, observable steps; preserve approvals, exception queues, audit logs, and rollback. Update `PIPELINE_STATE.md` after each completed stage. Never mark a stage complete because a document exists; require the stage exit evidence defined in the workflow. ## Enforce five gates Read [references/validation-gates.md](references/validation-gates.md) before assigning a readiness level. - `G1 Problem`: the target customer, painful job, current behavior, and willingness to act are supported. - `G2 Offer & Economics`: the offer is understandable, deliverable, priced, capacity-aware, and viable under stated assumptions. - `G3 Acquisition & Sale`: at least one channel and sales path show attributable progression toward commitment or payment. - `G4 Delivery & Retention`: value can be delivered consistently and the next-purchase, renewal, referral, or deliberate exit path is measured. - `G5 Automation & Scale`: the loop is observable, exceptions are controlled, economics remain acceptable, and automation does not outrun evidence. Report each gate as `passed`, `partial`, `failed`, or `not-tested` with evidence, sample size, period, limitations, and next falsifier. Apply these caps: - G1 failed: do not finalize an offer or acquisition plan as validated. - G2 failed: do not recommend paid acquisition or scale. - G3 failed: do not forecast repeatable revenue. - G4 failed or not-tested: do not claim a closed loop or sustainable growth. - G5 failed or not-tested: do not recommend autonomous scaling. - Any critical gate below `passed`: cap overall readiness below `scale-ready`. ## Apply evidence and metric discipline Read [references/metrics-and-economics.md](references/metrics-and-economics.md) before calculating economics or selecting metrics. Read [references/experimentation-standard.md](references/experimentation-standard.md) before designing or interpreting tests. Use observed events where available. Label estimates and benchmarks with source, date, geography, business model, and confidence. Never turn vanity metrics, verbal enthusiasm, unsigned proposals, or gross revenue into proof of profitability. Keep these distinctions explicit: - lead, qualified lead, opportunity, proposal, commitment, payment, activation, outcome, repeat purchase, renewal, expansion, referral, and churn; - booked revenue, collected cash, recognized revenue, gross margin, contribution margin, and operating profit; - channel-level CAC, blended CAC, payback, capacity constraint, and cash-conversion timing; - customer-level conversion and cohort-level retention. Never invent baselines, conversion rates, prices, costs, legal permissions, customer evidence, or completed experiments. When no data exists, build an instrumented test plan instead of a forecast dressed as fact. ## Design the operating system Read [references/funnel-sales-delivery.md](references/funnel-sales-delivery.md) when constructing acquisition, sales, CRM, payment, delivery, retention, or referral flows. Read [references/automation-governance.md](references/automation-governance.md) before adding tools, agents, integrations, or autonomous actions. Every handoff must define: - entry criteria and required fields; - owner and due time; - action and output; - exit criteria and next state; - exception path, stop rule, and audit record. Do not hide a broken business process behind more content, outreach, discounts, software, or agents. Diagnose the binding constraint first. Require explicit confirmation before sending messages, publishing content, launching ads, changing live prices, signing or accepting terms, charging or refunding money, purchasing tools, modifying production systems, or making another consequential external change. ## Initialize and validate a pack Initialize a reusable pack with: ```bash python3 scripts/init_loop.py --business "" --mode design ``` Validate structure, record integrity, formulas, stage state, and gate consistency with: ```bash python3 scripts/validate_loop.py python3 scripts/run_economics_tests.py ``` The initializer copies the canonical [scope](assets/scope.template.md), [pipeline state](assets/pipeline-state.template.md), [blueprint](assets/monetization-blueprint.template.md), [assumption](assets/assumption-ledger.template.csv), [customer evidence](assets/customer-evidence.template.csv), [offer](assets/offer-catalog.template.csv), [economics](assets/unit-economics.template.json), [experiment](assets/experiment-register.template.csv), [funnel](assets/funnel-ledger.template.csv), [cash](assets/cash-ledger.template.csv), [delivery](assets/delivery-scorecard.template.csv), [automation](assets/automation-register.template.csv), [metrics](assets/metrics-dashboard.template.md), [gate](assets/gate-report.template.md), and [validation](assets/validation-report.template.md) templates without overwriting existing files. Use [assets/evaluation-cases.template.json](assets/evaluation-cases.template.json) only when extending behavioral tests. Static checks do not prove demand, sales, delivery, retention, or profitability. Prefer forward tests with real but bounded customer and operational data. Mark simulations as `simulation`, self-runs as `self-test-fallback`, and unavailable validation as `not-tested`. ## Deliver the contract Read [references/output-contract.md](references/output-contract.md) before final delivery. Read [references/platform-compatibility.md](references/platform-compatibility.md) before installing or adapting this Skill for another Agent host. Deliver: - the bounded monetization thesis and current evidence cutoff; - the end-to-end loop with owners, states, handoffs, and stop rules; - customer, offer, price, economics, channel, sales, delivery, retention, and automation decisions; - the assumption ledger, experiment backlog, funnel definitions, and measurement plan; - G1–G5 status and overall readiness; - the next highest-information action, not a generic task list; - executed, manual, failed, simulated, and untested checks; - explicit unknowns, risks, approvals, and refresh triggers. Read [examples/sample-requests.md](examples/sample-requests.md) only when designing trigger tests or resolving scope confusion. Use [tests.json](tests.json) when changing this meta Skill's trigger, routing, or gate logic. ## Stop conditions Stop and request direction when the customer or offer boundary remains decision-changing ambiguous; decisive private data is inaccessible; legal, safety, payment, tax, privacy, or platform constraints are unresolved; losses could exceed the approved budget; evidence conflicts on a critical assumption; capacity cannot support promised delivery; or the user asks for fabricated proof, deceptive outreach, spam, guaranteed income, evasion, or unauthorized external action. For medical, legal, regulated financial, safety-critical, employment, insurance, controlled-goods, or other high-consequence businesses, provide planning and analysis only until current jurisdiction-specific professional review is documented. ## License Distribute and modify this Skill under the [MIT License](LICENSE.txt). ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [shangdizhiyan](https://github.com/shangdizhiyan) - **Source:** [shangdizhiyan/Vic-distill-industry-skills](https://github.com/shangdizhiyan/Vic-distill-industry-skills) - **License:** MIT Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-shangdizhiyan-vic-distill-industry-skills-build-monetization-loop - Seller: https://agentstack.voostack.com/s/shangdizhiyan - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.