AgentSkillsCN

designing-growth-loops

设计增长回路（如病毒式传播、推荐引流、获客回路，以及飞轮效应），并产出增长回路设计方案包（包括回路地图、回路评分卡、渠道适配与付费回路可行性分析、实验待办清单、效果衡量方案）。适用于增长团队打造全新增长回路，或在增量优化之外寻求创新突破。

SKILL.md
--- frontmatter
name: "designing-growth-loops"
description: "Design growth loops (viral/referral/acquisition loops, flywheels) and produce a Growth Loop Design Pack (loop map, loop scorecard, channel fit + paid-loop feasibility, experiment backlog, measurement plan). Use for growth teams creating new growth loops or innovating beyond incremental optimization."

Designing Growth Loops

Scope

Covers

  • •Turning a growth goal into a loop-based growth model (micro loops + macro loops)
  • •Designing and documenting loops: viral/referral, content/UGC, SEO, partner/integration, sales-assisted, and paid acquisition loops
  • •Choosing channels using a Customer × Business × Medium fit check
  • •Validating paid loops with unit economics (LTV, CAC, payback) gating
  • •Producing an actionable loop plan: loop map → scorecard → experiments → measurement

When to use

  • •“Design a growth loop / viral loop / referral loop”
  • •“Create a growth flywheel for <product>”
  • •“Map our micro + macro growth loops and prioritize which to build”
  • •“We need new growth loops (not just optimize ads/onboarding)”
  • •“Decide whether a paid acquisition loop is viable”

When NOT to use

  • •You haven’t clarified the ICP/problem or value proposition (use problem-definition).
  • •You’re still establishing PMF and need a PMF signal set (use measuring-product-market-fit).
  • •You only need an experiment list/prioritization, not loop design (use prioritizing-roadmap).
  • •You’re making a one-way-door launch decision (use shipping-products / running-decision-processes).

Inputs

Minimum required

  • •Product + target user/ICP (and 1–2 key segments)
  • •Current stage (pre-PMF / early PMF / growth / mature) and current primary growth channel(s)
  • •The core value moment (what users do when they “get value”)
  • •A baseline snapshot of the growth system (best available): acquisition sources, conversion funnel, retention/engagement, referrals/sharing
  • •Constraints: budget, timebox, brand/safety, platform policy, legal/privacy, engineering capacity
  • •For paid loops: rough unit economics (LTV, gross margin, CAC/payback targets) or a proxy

Missing-info strategy

  • •Ask up to 5 questions from references/INTAKE.md, then proceed.
  • •If data is missing, proceed with explicit assumptions and label confidence.
  • •Do not request secrets or PII; prefer aggregated metrics or redacted excerpts.

Outputs (deliverables)

Produce a Growth Loop Design Pack (Markdown in-chat; or as files if requested) containing:

  1. •Context snapshot (goal, ICP/segments, constraints, timebox)
  2. •Loop inventory + baseline (current loops and where the system currently gets growth)
  3. •Loop map (qualitative model) (micro loops + macro loop; how loops connect)
  4. •Loop candidates + mechanism library (platform/channel mechanisms; ethical/policy-compliant)
  5. •Loop scorecard + selection (top 1–2 loops to build/scale; optimize vs innovate recommendation)
  6. •Measurement plan (loop KPIs, leading indicators, required instrumentation)
  7. •Experiment backlog + 30/60/90 plan (tests, sequencing, dependencies, owners if known)
  8. •Risks / Open questions / Next steps (always included)

Templates and checklists:

Workflow (7 steps)

1) Intake + growth goal framing

  • •Inputs: User context; references/INTAKE.md.
  • •Actions: Clarify the growth goal (what metric, by when), the target segment(s), and constraints (budget, brand, platform rules, capacity). Decide whether the priority is innovation (new loop) vs optimization (existing loop).
  • •Outputs: Context snapshot + “decision this work informs.”
  • •Checks: A stakeholder can answer: “Which metric changes by when, and what will we do differently if it doesn’t?”

2) Baseline the current growth system (loops + funnel)

  • •Inputs: Current acquisition sources, funnel, retention, referral/share, unit economics (if any).
  • •Actions: Inventory existing loops (even if weak). Identify the core value moment and the “loop output” that could feed back (invites, content, word-of-mouth, spend, integrations).
  • •Outputs: Loop inventory + baseline table.
  • •Checks: Baseline includes at least one number for each: acquisition volume, activation rate, retention/engagement proxy.

3) Generate loop candidates (micro + macro)

  • •Inputs: Baseline + constraints.
  • •Actions: Create 6–10 loop hypotheses across categories (viral/referral, content/UGC, SEO, partner/integration, sales, paid). For each, specify: input → action → output → feedback. Include at least one “bigger bet” loop if in a fast-moving category.
  • •Outputs: Loop candidates list + draft mechanism library.
  • •Checks: Each candidate has a plausible “self-reinforcing” feedback path and a likely cycle time.

4) Model loops qualitatively (shared understanding)

  • •Inputs: Loop candidates; stakeholder context.
  • •Actions: Produce a qualitative loop map: micro loops connected into a macro loop. Document assumptions, bottlenecks, and where you expect compounding.
  • •Outputs: Loop map (diagram or table) + bottleneck hypotheses.
  • •Checks: Someone unfamiliar with the product can explain “how we grow” in 60 seconds using the map.

5) Quantify + prioritize (scorecard + gates)

  • •Inputs: Qual loop map; best-available metrics.
  • •Actions: Estimate loop throughput with simple math (conversion × frequency × invites/content × acceptance). Score loops using a scorecard (impact, confidence, effort, cycle time). Apply gates:
    • •Paid loops: only proceed if LTV/margin supports CAC/payback targets.
    • •Channel fit: ensure Customer × Business × Medium alignment.
  • •Outputs: Loop scorecard + top 1–2 loop picks + innovate/optimize split recommendation.
  • •Checks: Each chosen loop has (a) a measurable KPI, (b) a first experiment, and (c) a reason it wins vs alternatives.

6) Design the measurement plan (metrics + instrumentation)

  • •Inputs: Selected loop(s).
  • •Actions: Define loop KPIs and leading indicators; specify required events/properties and dashboards. Identify instrumentation gaps and the minimum tracking needed to learn.
  • •Outputs: Measurement + instrumentation plan.
  • •Checks: Every experiment metric is traceable to an event definition and a data source.

7) Build the experiment plan + quality gate

  • •Inputs: Draft pack; references/CHECKLISTS.md and references/RUBRIC.md.
  • •Actions: Create an experiment backlog and 30/60/90 plan (sequencing, dependencies, owners if known). Run the checklist and score with the rubric. Always include Risks / Open questions / Next steps.
  • •Outputs: Final Growth Loop Design Pack.
  • •Checks: Next 2 weeks of work are unblocked and measurable; risks include policy/ethics considerations.

Quality gate (required)

Examples

Example 1 (B2B SaaS, partner/integration loop):
“Use designing-growth-loops. Product: AI onboarding assistant for mid-market HR teams. Goal: +30% WAU in 90 days. Channels today: outbound + partnerships. Output: a Growth Loop Design Pack with an integration/partner loop and a referral loop, including metrics and a 30/60/90 experiment plan.”

Example 2 (B2C, viral/content loop):
“We’re building a mobile photo editor for creators. Goal: grow from 20k to 60k MAU in 8 weeks. Output a loop map, a mechanism library for Instagram/TikTok sharing, and a prioritized experiment backlog.”

Boundary example (not a loop problem):
“Write copy for our landing page headline.”
Response: this is primarily copywriting/positioning, not loop design; clarify the goal and use copywriting or a messaging skill instead.