Growth Loop Builder

Growth Loop Builder

What a Growth Loop Is

A growth loop is a system where growth feeds itself. Each user, action, or dollar of revenue creates the input for acquiring the next user — without proportional increase in effort or spend.

The opposite of a growth loop: a funnel. Funnels require constant top-of-funnel effort to produce output. When you stop feeding the funnel, growth stops. Growth loops compound — each cycle produces slightly more than the last.

The test: "Does each new user create the conditions for the next new user?" If yes, you have a loop. If not, you have a funnel.

Quick Start

Human-First Response

Default: answer in chat first with Verdict, Do this now, and Details saved. Keep chat under 150 words unless asked; do not paste the full artifact. Write/update the requested .md file as the durable record for agents.

Say: "Build my growth loop" or "How do I grow without paid ads?" or "Design a growth loop for me"

Output: growth-loop.md — the highest-potential loop type, specific design, implementation steps, and the metric that tells you the loop is working.

---

Parallel Execution

Four loop types. Four independent evaluations. All run simultaneously.

Before spawning, gather:

• founder-context.md — product description, customer profile, current acquisition channels
• customer-profile.md — how the target customer makes decisions, who they influence
• pmf-assessment.md — which signals are strongest (points to which loop type fits best)
• Any data on current user behavior: referrals, content creation, usage patterns

Spawn these 4 agents simultaneously:

Agent 1 — Viral Loop Evaluator A viral loop exists when users invite others as a natural part of using the product.

True viral loops:

• Calendly: sending a booking link exposes the product to a new person
• Dropbox: sharing a folder requires the recipient to have Dropbox
• Loom: watching a loom video shows the product to the viewer
• Zoom: joining a call requires downloading Zoom

Questions to assess fit:

• Does using the product naturally expose it to non-users?
• Is there a feature where value increases when others join?
• Can the invite/share be a core product action rather than a marketing bolt-on?
• What is the current viral coefficient (k-factor)? (invites sent per user × conversion rate)

Returns: Viral loop fit score (1–5) + whether a natural viral mechanic exists + specific design + k-factor estimate + what would need to be true for k > 1

Agent 2 — Content Loop Evaluator A content loop exists when users generate content that attracts new users, who generate more content.

Content loop examples:

• YouTube: creators upload → viewers discover → some become creators
• Yelp: reviewers write → searchers find → some become reviewers
• Reddit: posters share → readers engage → some become posters
• Product Hunt: makers launch → hunters discover → some become makers

Questions to assess fit:

• Does using the product generate content or data that's valuable to non-users?
• Can that content be indexed by search engines or discovered on social platforms?
• Is there a "creator" user type that produces content consumed by a "consumer" type?
• What's the incentive for users to create content? (recognition, SEO benefit, social proof)

Returns: Content loop fit score (1–5) + whether a content creation mechanic exists + specific design + SEO/discovery opportunity + what type of content users already create that could be amplified

Agent 3 — Product Loop Evaluator A product loop exists when usage of the product generates data or network effects that improve the product, attracting more users.

Product loop examples:

• Waze: drivers report traffic → map improves → attracts more drivers → more data
• Spotify: listeners create playlists → algorithm improves → attracts more listeners
• Duolingo: learners practice → AI improves → better app → more learners
• Notion: users create templates → template gallery grows → attracts new users

Questions to assess fit:

• Does more usage generate data that improves the product for all users?
• Is there a "community asset" (template gallery, benchmark data, community answers) that grows with usage?
• Do network effects exist — does the product get more valuable as more users join?
• What unique data asset is the product generating that competitors don't have?

Returns: Product loop fit score (1–5) + whether data/network effects exist + specific design + the unique data asset being generated + competitive moat this creates over time

Agent 4 — Sales Loop Evaluator A sales loop exists when revenue from customers is reinvested into acquisition in a way that generates more revenue than it costs — the loop funds itself.

Sales loop examples:

• HubSpot: content → leads → revenue → more content team → more leads
• Salesforce: revenue → enterprise sales team → bigger contracts → more revenue
• Any SaaS with < 12 month payback period: CAC paid back → reinvest in acquisition → more customers

Questions to assess fit:

• What is the current (estimated) customer acquisition cost?
• What is the estimated LTV / payback period?
• If payback period is < 12 months, is there budget to reinvest in the loop?
• What acquisition channel has the best CAC and is most scalable?

Returns: Sales loop fit score (1–5) + estimated CAC/LTV ratio + whether payback period supports a self-funding loop + specific channel to invest in + when this loop makes sense (usually post-PMF, not pre-PMF)

Wait for all 4 agents. Orchestrator selects and designs.

---

Synthesis (Orchestrator Only)

1. Rank loops by fit score. The highest-scoring loop gets the full design treatment.

2. Reality check: The best loop type is the one that works with the product's natural mechanics — not the one that sounds most exciting. A marketplace forcing a viral loop that doesn't fit will fail. A B2B SaaS forcing a content loop without writers will fail.

3. Design the primary loop in full detail:

[Input] → [Product Action] → [Output] → [Acquisition Channel] → [New User] → [Input]

4. Write growth-loop.md:

# Growth Loop — [Product Name] — [YYYY-MM-DD]

## Primary Loop: [Viral / Content / Product / Sales]

### The Loop

[Input] → [Action] → [Output] → [Channel] → [New User joins] → [becomes Input]

### Why This Loop, Not the Others
| Loop Type | Fit Score | Reason Not Primary |
|-----------|-----------|-------------------|
| Viral | [X]/5 | [One sentence] |
| Content | [X]/5 | [One sentence] |
| Product | [X]/5 | [One sentence] |
| Sales | [X]/5 | [One sentence] |

### The Design

**Step 1 — [Input]:** [What triggers the loop — a user action, a piece of content, a referral]

**Step 2 — [Product Action]:** [What happens in the product that creates the output]

**Step 3 — [Output]:** [What the loop produces that attracts new users]

**Step 4 — [Channel]:** [How the output reaches potential new users]

**Step 5 — [New User]:** [What the new user does that feeds back into Step 1]

### Implementation Plan

**Week 1:** [The minimum change to create a working version of this loop]
**Week 2–3:** [What makes the loop more efficient]
**Week 4+:** [What scales the loop once it's working]

### The Loop Metric
The single number that tells you the loop is working:
**[Metric]** — Target: [X]. Current: [Y].

A loop is working when this metric is positive and improving cycle-over-cycle.

### What Could Break It
[The one assumption in this loop that must be true for it to work — and how to test it]

---

Sequential Fallback (Codex)

Evaluate each loop type sequentially:

1. Viral loop → fit score + design notes
2. Content loop → fit score + design notes
3. Product loop → fit score + design notes
4. Sales loop → fit score + design notes
5. Select highest fit → full design → write growth-loop.md

---

Related Skills

• Use pmf-signal-reader before this — word-of-mouth signal points to viral or content loops; engagement depth points to product loops
• Use north-star-definer — the loop metric should align with the north star
• Use retention-loop-designer alongside — retention and growth loops reinforce each other
• Use co-founder to decide when to focus on growth vs. when to keep compounding on product quality

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: gvkhosla
• Source: gvkhosla/founder-skills
• License: MIT
• Homepage: https://fskills.xyz

Install and usage instructions live in the source repository linked above.