---
name: lean-startup-method
description: Apply the Build-Measure-Learn feedback loop to validate business hypotheses with minimum viable products. Use when the user asks about lean startup, MVP design, validated learning, build-measure-learn, or whether to pivot or persevere.
triggers:
  - "lean startup"
  - "MVP"
  - "validated learning"
  - "build measure learn"
  - "pivot or persevere"
disable-model-invocation: false
---

# Lean Startup Method

## Intent
Use Eric Ries's Build-Measure-Learn loop to systematically test business hypotheses and determine whether to pivot or persevere. Apply this skill when launching a new product, entering a new market, or facing uncertainty about customer demand.

## Inputs
- Business idea or hypothesis statement
- Target customer segment definition
- Key assumptions (value hypothesis, growth hypothesis)
- Available resources and timeline
- Current stage of product development

## Process (SOP)
1. Step 1: Identify the riskiest assumption underlying your business idea (often the value hypothesis — will customers actually use/buy this?).
2. Step 2: Define a Minimum Viable Product (MVP) that tests that single assumption with minimum effort — strip every feature that does not directly serve the learning goal.
3. Step 3: Define the metric that will measure success before building anything (e.g., conversion rate, retention cohort, willingness-to-pay signal).
4. Step 4: Build the MVP in the shortest possible cycle — days or weeks, not months.
5. Step 5: Deploy the MVP to a small, representative cohort of real customers (not friends and family).
6. Step 6: Measure actual behavior against the pre-registered metric using innovation accounting (cohort-based analysis, not vanity metrics).
7. Step 7: Analyze results to determine whether the hypothesis is validated, invalidated, or inconclusive.
8. Step 8: Decide: pivot (change one or more hypotheses while keeping learned insights) or persevere (continue optimizing the same fundamental strategy).
9. Step 9: Document what was learned in a structured format — hypothesis → test → result → decision — so knowledge compounds across iterations.
10. Step 10: Repeat the loop, progressively increasing investment only as validation accumulates.

## Outputs
- A tested MVP deployed to real customers
- A clear pivot-or-persevere decision backed by cohort data
- A documented learning log of validated and invalidated hypotheses
- Reduced waste from building features nobody wants

## Evidence
Source: "The Lean Startup" by Eric Ries. Key concepts: Build-Measure-Learn loop, MVP definition, validated learning, innovation accounting, pivot vs. persevere, vanity metrics vs. actionable metrics, cohort analysis.

## Examples
A SaaS startup wants to test demand for an AI-powered invoicing tool. They skip the full product and build a concierge MVP: manually process invoices for 10 small businesses behind a single landing page. When 7 of 10 sign up for the paid waitlist, the value hypothesis is validated and they proceed to build the product.

## Metrics
- Time per Build-Measure-Learn cycle (target: under 2 weeks)
- Number of hypotheses validated vs. invalidated per quarter
- Customer retention cohort curves (flattening or declining)
- Percentage of roadmap driven by validated learning vs. speculation

## Anti-Patterns
- Building a "minimum" product that is actually bloated with unnecessary features
- Measuring vanity metrics (total signups, page views) instead of actionable cohort metrics
- Pivoting too early based on inconclusive data from too few customers
- Pivoting too late, refusing to abandon a failing strategy despite evidence
- Calling a pivot a failure instead of a strategic redirection
- Running multiple concurrent experiments without isolating variables
