Reforge — Growth Models
Summary
A growth model answers one question: how does your product grow? The model must be built from the perspective of users (not capital), grounded in loops (not funnels), and kept simple enough to actually use. The core mental model is a three-tier modeling system — qualitative, loop-specific quantitative, and end-to-end quantitative — each serving a different purpose.
Why Most Growth Models Fail
Three common mistakes that render models useless:
- Wrong perspective — built around inflow/outflow of capital rather than users or customers.
- Wrong law — built on funnels rather than loops, which breaks the compounding logic.
- Pursuit of perfection — too detailed and complicated to use on a day-to-day or week-to-week basis.
Three Types of Growth Models
| Model | Best For | Limitation |
|---|---|---|
| Qualitative model | Communication, brainstorming, strategic alignment | Lacks accuracy for prioritization and goal setting |
| Loop-specific quant model | Validating and prioritizing ideas within a single loop | Cannot show how multiple loops interact |
| End-to-end quant model | Goal setting, big strategic bets, directional predictions | Too complex for broad team communication |
The right answer is to use all three for different purposes — not to pick one.
Building a Qualitative Growth Model
The qualitative model has three goals: (1) create an easy-to-communicate tool, (2) align team understanding and language, (3) identify and evaluate the strategic picture.
Four steps to build it:
- Define the output — what does the model produce? (Active users, revenue, etc.)
- Add habit/engagement loops — how do existing users keep coming back? Three trigger types: user-triggered, company-triggered, environment-triggered. See 06-reference/2026-04-03-reforge-engagement-activation for engagement loop mechanics.
- Add acquisition loops — how does one cohort lead to the next? Map the specific loop types from 06-reference/2026-04-03-reforge-acquisition-loops.
- Add linear channels — what non-looping activities feed the system?
Linear channels play two roles in the model: (a) feeding loops that are not self-sustaining, and (b) providing activation energy to get new loops spinning.
Growth Model and the Four Pillars
The growth model connects to the four strategic pillars from 06-reference/2026-04-03-reforge-defining-strategy:
- How do we acquire? — Acquisition loops and linear channels.
- How do we retain? — Engagement/habit loops and retention mechanics. See 06-reference/2026-04-03-reforge-retention-is-the-output and 06-reference/2026-04-03-reforge-retention-measurement.
- How do we monetize? — The monetization model that enables or constrains the loops. See 06-reference/2026-04-03-reforge-monetization-strategy.
- How do we defend and improve? — See 06-reference/2026-04-03-reforge-monetization-defensibility.
A growth loop is a closed system: input (new or returning user) -> series of actions -> output that is directly reinvested as input. This definition applies to both acquisition loops and engagement loops.
Relevance to projects:
- 01-projects/data-marketplace/index — The qualitative growth model should be the first artifact built. Map the output (active data consumers), the engagement loop (what triggers repeat queries), the acquisition loops (content? sales? viral sharing of datasets?), and the linear channels (outbound, events, partnerships). This becomes the shared strategic language.
- 01-projects/newsletter/index — The newsletter likely appears in the qualitative model as a linear channel feeding acquisition loops for other projects. The question is whether it can become its own engagement loop with a company-triggered cadence.
Connects to 06-reference/2026-04-03-reforge-defining-strategy (growth model as the bridge between mission and execution), 06-reference/2026-04-03-reforge-acquisition-loops (the acquisition layer of the model), 06-reference/2026-04-03-growth-loops-new-funnels (why loops replace funnels in the model), and 06-reference/2026-04-03-racecar-growth-framework (another visualization of the same growth system).
Open Questions
- Has Ray Data Co built a qualitative growth model for any of its projects yet? This is the highest-leverage first step — it takes an hour and aligns all subsequent decisions.
- For the data marketplace, is the engagement loop user-triggered (they have a recurring data need), company-triggered (we push new datasets or alerts), or environment-triggered (regulatory changes, market events create demand)?
- At what point does it make sense to invest in an end-to-end quantitative model? Probably not until there is enough data to populate it — which means after launch with real users.