designing-growth-loops

Original source / Raw SKILL.md

---
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](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:
- [references/TEMPLATES.md](references/TEMPLATES.md)
- [references/CHECKLISTS.md](references/CHECKLISTS.md)
- [references/RUBRIC.md](references/RUBRIC.md)
- Expanded guidance: [references/WORKFLOW.md](references/WORKFLOW.md)

## Workflow (7 steps)

### 1) Intake + growth goal framing
- **Inputs:** User context; [references/INTAKE.md](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](references/CHECKLISTS.md) and [references/RUBRIC.md](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)
- Use [references/CHECKLISTS.md](references/CHECKLISTS.md) and [references/RUBRIC.md](references/RUBRIC.md).
- Always include: **Risks**, **Open questions**, **Next steps**.

## 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.



---

## Referenced Files

> The following files are referenced in this skill and included for context.

### references/INTAKE.md

```markdown
# Intake (Question Bank)

Ask up to 5 questions at a time. If answers are missing, proceed with explicit assumptions and label confidence.

## A) Goal + constraints (required)
1) What is the growth goal (metric + target + date)? (e.g., WAU, revenue, activated users)
2) Which segment(s) matter most right now (ICP, persona, geo, plan tier)?
3) What are the key constraints: budget, team capacity, brand risk, platform policies, legal/privacy, timebox?
4) Are we optimizing an existing loop, or do we need a new loop? (If unclear, default to a balanced plan.)

## B) Product + value moment (required)
5) What is the “core value moment” (the action that means the user got value)?
6) What is the primary user job-to-be-done / use case?
7) What are the top reasons users churn or don’t activate (if known)?

## C) Baseline growth system (best available)
8) What currently drives acquisition? (top 3 sources + rough volumes)
9) What does the funnel look like? (visit → signup → activation → retained)
10) What retention/engagement proxy do you have (e.g., D7/D30, WAU/MAU, repeat purchase)?
11) Do users share/invite today? If yes, how often and through what mechanism?

## D) Unit economics (needed for paid loops)
12) Rough LTV (or ARPA × gross margin × expected lifetime) and CAC today (if any)?
13) Any payback constraints? (e.g., 3 months, 12 months)
14) Gross margin and variable costs that affect paid viability?

## E) Channels + distribution context
15) Which channels have you tried and what happened?
16) Where does your audience already spend attention? (communities, platforms, search, partners)
17) Any channel-specific strengths/weaknesses? (e.g., strong content, strong product virality, strong sales motion)

## F) Measurement + tooling
18) What analytics tooling exists today (events, attribution, dashboards)?
19) Any known instrumentation gaps that make measurement hard?

## Data handling rules
- Do not request secrets (API keys, tokens).
- If the user data includes PII, ask for aggregated metrics or redacted excerpts.


```

### references/TEMPLATES.md

```markdown
# Templates (Copy/Paste)

Use these templates to produce a **Growth Loop Design Pack**. If writing files, keep everything under a user-specified folder (e.g., `docs/growth-loops/`).

---

## 1) Context snapshot

**Product:**  
**Stage:** pre-PMF / early PMF / growth / mature  
**Target segment(s):**  
**Growth goal (metric + target + date):**  
**Decision this informs:**  
**Current primary channels:**  
**Constraints (budget/capacity/brand/policy/legal/privacy):**  
**Core value moment (definition):**  

---

## 2) Loop inventory + baseline

### Baseline (best available)
| Area | Metric / definition | Current | Source | Notes / confidence |
|---|---|---:|---|---|
| Acquisition volume |  |  |  |  |
| Activation rate |  |  |  |  |
| Retention/engagement proxy |  |  |  |  |
| Referral/share |  |  |  |  |
| Unit economics (if known) | LTV, margin, CAC, payback |  |  |  |

### Existing loops (if any)
| Loop name | Type | Input → Action → Output → Feedback | Strength today (H/M/L) | Evidence |
|---|---|---|---|---|
|  |  |  |  |  |

---

## 3) Loop hypothesis card (for candidates)

**Loop name:**  
**Type:** viral/referral / content/UGC / SEO / partner/integration / sales / paid  
**Target segment:**  
**Core mechanism (1 sentence):**  

### Loop model
- **Input:** (what starts the loop?)
- **Action:** (what users/teams do)
- **Output:** (what gets produced: invites, content, spend, integrations)
- **Feedback:** (how output creates new input)
- **Cycle time:** (how long for one loop turn?)

### Preconditions
- Product prerequisites:
- Operational prerequisites:
- Risk/policy constraints:

### Success definition
- **Loop KPI:**  
- **Leading indicators (early):**  
- **Lagging indicator (business):**  

---

## 4) Loop map (qualitative model)

### Micro loops → macro loop table
| Micro loop | Input | Action | Output | Feedback path | Primary bottleneck |
|---|---|---|---|---|---|
|  |  |  |  |  |  |

### Optional Mermaid diagram (if supported)
```mermaid
flowchart LR
  A[Acquisition input] --> B[Activation/value moment]
  B --> C[Loop output: invites/content/$$]
  C --> A
```

---

## 5) Channel fit triad (Customer × Business × Medium)

| Candidate channel | Customer need/context | Business goal fit | Medium strength match (audio/video/text/interactive) | Verdict (Go/No-go) | Notes |
|---|---|---|---|---|---|
|  |  |  |  |  |  |

---

## 6) Paid loop feasibility gate

Use this section if recommending paid acquisition.

**Known/estimated:**
- LTV (gross):  
- Gross margin %:  
- Target CAC:  
- Payback target (months):  

**Gate checks (directional):**
- LTV × margin supports target CAC (and payback window is realistic)
- Retention is strong enough that CAC won’t be “wasted”
- Attribution and conversion tracking are good enough to learn

**Verdict:** viable / not yet viable / needs more data  
**If not yet viable, prerequisites:** (e.g., retention, monetization, onboarding, measurement)

---

## 7) Loop scorecard (pick top 1–2)

| Loop candidate | Impact | Confidence | Effort | Cycle time | Risk | Notes | Total |
|---|---:|---:|---:|---:|---:|---|---:|
|  |  |  |  |  |  |  |  |

Scoring tip: use 1–5 per column, define what “5” means for your context.

---

## 8) Experiment backlog

| Priority | Experiment | Loop | Hypothesis (mechanism) | Metric (leading) | Metric (lagging) | Effort | Timebox | Dependencies |
|---:|---|---|---|---|---|---|---|---|
| 1 |  |  |  |  |  |  |  |  |

---

## 9) Measurement + instrumentation plan

### Metrics
| Loop | KPI (headline) | Leading indicators | Data source | Notes |
|---|---|---|---|---|
|  |  |  |  |  |

### Required events/properties
| Event name | When it fires | Properties | Used for |
|---|---|---|---|
|  |  |  |  |

---

## 10) 30/60/90 plan

### Next 30 days (de-risk + first tests)
-  

### 60 days (iterate toward a working loop)
-  

### 90 days (scale winners, cut losers)
-  

---

## 11) Risks / Open questions / Next steps

**Risks**
-  

**Open questions**
-  

**Next steps**
-  


```

### references/CHECKLISTS.md

```markdown
# Checklists (Quality Gate)

Use this checklist before finalizing a Growth Loop Design Pack.

## A) Scope + decision clarity
- [ ] The growth goal is explicit (metric + target + date).
- [ ] Target segment(s) are defined (not “everyone”).
- [ ] Constraints are stated (budget, capacity, brand risk, platform policies, legal/privacy).
- [ ] The decision this work informs is explicit (which loop(s) we invest in; whether paid is viable; innovate vs optimize split).

## B) Baseline is grounded
- [ ] Baseline includes at least one metric for acquisition, activation, and retention/engagement.
- [ ] Existing loops (if any) are documented with evidence.
- [ ] The core value moment is defined and consistent with the product’s value proposition.

## C) Loop model quality (qualitative)
- [ ] Micro loops are specified as input → action → output → feedback.
- [ ] A macro loop connects the micro loops (shared understanding of “how we grow”).
- [ ] Cycle time is estimated per loop (how long for one turn).
- [ ] Bottlenecks are named (what limits compounding today).

## D) Prioritization + gating
- [ ] Loop scorecard is present and picks top 1–2 loops (with rationale).
- [ ] Channel fit triad is applied (Customer × Business × Medium).
- [ ] Paid loop feasibility is gated by unit economics and measurement feasibility (if applicable).

## E) Measurement + learning plan
- [ ] Each selected loop has a KPI + leading indicators.
- [ ] Instrumentation requirements are specified (events/properties) or gaps are called out.
- [ ] Experiments have hypotheses, metrics, and timeboxes.

## F) Safety + ethics
- [ ] Mechanisms are policy-compliant and avoid manipulative/dark-pattern tactics.
- [ ] Risks include brand/trust and platform policy considerations.

## G) Required closing section
- [ ] The pack includes **Risks**, **Open questions**, and **Next steps**.


```

### references/RUBRIC.md

```markdown
# Rubric (Score 1–5)

Score the Growth Loop Design Pack across dimensions. Use scores to decide whether to execute as-is or iterate.

## 1) Decision usefulness
1 = No decision; mostly brainstorming  
3 = Decision is named but implications are fuzzy  
5 = Clear decision + clear “what we’ll do differently” based on outcomes

## 2) Loop clarity (micro + macro)
1 = Tactics list; no feedback paths  
3 = Some loops defined but weak connections/bottlenecks  
5 = Clear micro loops + coherent macro loop; bottlenecks explicit; explainable in 60 seconds

## 3) Evidence grounding
1 = No baseline metrics; no assumptions labeled  
3 = Baseline exists but is incomplete or confidence unclear  
5 = Baseline + assumptions + confidence are explicit; claims tied to evidence or labeled hypotheses

## 4) Prioritization quality
1 = No scorecard; “pick everything”  
3 = Scorecard exists but selection rationale is thin  
5 = Scorecard drives a clear top 1–2 selection with trade-offs and rationale

## 5) Unit economics + paid gating (if applicable)
1 = Recommends paid without LTV/CAC/payback reasoning  
3 = Mentions unit economics but no clear gate  
5 = Paid loop viability is explicitly gated; prerequisites listed if not viable

## 6) Measurement + instrumentation readiness
1 = Metrics vague; no instrumentation plan  
3 = Metrics defined but data sources/gaps unclear  
5 = KPI + leading indicators; required events/properties; gaps + fixes are explicit

## 7) Actionability
1 = Vague recommendations (“go viral”)  
3 = Some experiments but weak sequencing/dependencies  
5 = Prioritized backlog + 30/60/90 plan; next 2 weeks unblocked

## 8) Safety + ethics
1 = Encourages manipulative tactics / ignores policy risk  
3 = Mentions safety but weak application  
5 = Explicit policy/ethics constraints; risks and mitigations are concrete

### Interpretation (suggested)
- **30–40:** Strong; execute and review cadence weekly.
- **22–29:** Good directional; tighten weak areas before big bets.
- **<22:** Rework inputs/baseline/loop definitions; too speculative.


```

### references/WORKFLOW.md

```markdown
# Workflow (Expanded)

This file expands the steps from `../SKILL.md` with additional guidance, heuristics, and common pitfalls.

## Step 1 — Intake + goal framing
Anchor on a decision (what you’ll do differently), not curiosity.

Common decisions:
- Invest in **new loops** vs optimize existing channels
- Choose the **top 1–2 loops** for the next quarter
- Decide whether **paid acquisition** is viable (unit economics gate)

Pitfall: “We need growth” without a metric/timebox yields a generic brainstorming doc.

## Step 2 — Baseline the current growth system
Even if you’re “pre-loops,” you still have a system:
- a channel that brings users in
- a value moment
- some retention (even if poor)

Baseline should include:
- Acquisition: where users come from + rough volume
- Activation: what % hits the value moment
- Retention: a proxy (D7/D30, WAU/MAU, repeat purchase)
- Referral/share: any organic sharing/invites today

Pitfall: designing loops without knowing where the real bottleneck is (activation vs retention vs distribution).

## Step 3 — Generate loop candidates (micro + macro)
Treat loop design as hypothesis generation.

Loop categories (choose what fits your product):
- Viral/referral loop (invites, sharing, collaboration)
- Content/UGC loop (users create content that acquires users)
- SEO loop (content → search traffic → more content)
- Partner/integration loop (distribution via ecosystems)
- Sales loop (champions → expansion → internal referrals)
- Paid loop (spend → acquisition → revenue → reinvest)

Heuristic: in fast-moving categories, allocate meaningful effort to “bigger bet” loop innovation, not just incremental tuning.

Pitfall: listing tactics (“do TikTok”) without specifying the feedback mechanism (how outputs become future inputs).

## Step 4 — Model loops qualitatively (shared understanding)
You want a model that is:
- simple enough to explain
- explicit about assumptions
- clear about bottlenecks

Represent loops using:
- a table (input → action → output → feedback)
- and optionally a diagram (e.g., Mermaid) if your environment supports it

Pitfall: “funnel thinking” only—loops require a feedback path and cycle time.

## Step 5 — Quantify + prioritize (scorecard + gates)
Use best-available numbers to approximate throughput.

Simple loop math (example):
`new_users_per_week ≈ active_users × share_rate × invite_accept_rate`

Prioritization dimensions:
- Impact (if it works, how big)
- Confidence (evidence, precedent, prerequisites)
- Effort (engineering, content ops, sales ops)
- Cycle time (how fast you’ll learn / compound)
- Risk (policy, brand, trust)

Gates:
- **Paid loop gate:** if LTV/margins can’t support CAC/payback, do not recommend scaling paid; recommend fixing retention/monetization first.
- **Channel fit triad:** customer need × business goal × medium strength must align (don’t force audio if your product needs visuals).

Pitfall: choosing a loop that looks exciting but has a 6–12 month cycle time when you need learning in 2–4 weeks.

## Step 6 — Measurement plan (metrics + instrumentation)
A loop is only real if it’s measurable.

Define:
- Loop KPI (one headline measure per loop)
- Leading indicators (early signals)
- Required instrumentation (events + properties)
- Attribution assumptions (what you can/can’t infer)

Pitfall: shipping experiments without event definitions; you can’t learn and the loop “dies in analytics.”

## Step 7 — Experiment plan + quality gate
Translate loops into a practical plan:
- 30 days: de-risk prerequisites + first tests
- 60 days: iterate toward a working version
- 90 days: scale the winners + cut losers

Quality gate reminders:
- Run [CHECKLISTS.md](CHECKLISTS.md)
- Score with [RUBRIC.md](RUBRIC.md)
- Always include **Risks / Open questions / Next steps**


```