Lean Startup

Eric Ries' Lean Startup methodology emphasizes rapid experimentation and validated learning. The Build-Measure-Learn feedback loop is the core engine for turning ideas into sustainable businesses.

The Flow

Key Principles

1. Validated Learning: Progress through scientific testing of hypotheses
2. Build-Measure-Learn: Rapid feedback loops
3. Minimum Viable Product (MVP): Learn with least effort
4. Pivot or Persevere: Data-driven decision to change direction
5. Innovation Accounting: Measure progress in uncertain conditions

VisionSpec Mapping

Lean Artifact	VisionSpec Type	Purpose
Hypothesis Document	MRD	Assumptions to validate
MVP PRD	PRD	Minimum feature set for learning
Experiment Design	UXD	Test methodology
Pivot/Persevere Decision	Narrative	Learning documentation

Using the Lean Startup Profile

Initialize a Project

multispec init my-hypothesis --profile lean-startup

Create Hypothesis Document (MRD)

multispec draft mrd -p my-hypothesis

The Hypothesis Document template includes:

• Problem hypothesis
• Solution hypothesis
• Customer segment hypothesis
• Riskiest assumptions (ordered by risk)
• Leap of faith assumptions
• Success/failure criteria

Create MVP PRD

multispec draft prd -p my-hypothesis

The MVP PRD template includes:

• Learning objectives (what we want to learn)
• MVP scope (minimum features)
• What's explicitly excluded
• Build vs. fake (Wizard of Oz options)
• Timeline constraints

Create Experiment Design (UXD)

multispec draft uxd -p my-hypothesis

The Experiment Design template includes:

• Hypothesis being tested
• Experiment type (A/B, landing page, concierge, etc.)
• Success metrics (actionable, not vanity)
• Sample size and duration
• Data collection method
• Analysis plan

Document Pivot/Persevere Decision

multispec draft narrative-1p -p my-hypothesis

The Pivot Document includes:

• What we learned
• Data supporting conclusions
• Pivot type (if applicable)
• New hypotheses to test
• Resources needed

Rubric Categories

Hypothesis Document Evaluation (MRD)

Category	Weight	Description
Hypothesis Clarity	25%	Clear, falsifiable statements
Assumption Identification	20%	Key assumptions explicit
Risk Prioritization	20%	Riskiest assumptions first
Testability	15%	Hypotheses can be tested
Success Criteria	15%	Clear pass/fail thresholds
Customer Definition	5%	Specific customer segment

MVP PRD Evaluation

Category	Weight	Description
Learning Focus	25%	Learning objectives clear
Minimum Scope	25%	Truly minimum for learning
Exclusions Clear	15%	What's NOT included
Build vs. Fake	15%	Considered Wizard of Oz
Time-Boxed	10%	Iteration has deadline
Measurable	10%	Can measure outcomes

Experiment Design Evaluation (UXD)

Category	Weight	Description
Hypothesis Clarity	20%	What exactly are we testing
Actionable Metrics	20%	Not vanity metrics
Experiment Validity	20%	Will results be meaningful
Sample Size	15%	Statistically valid
Data Collection	15%	How we'll gather data
Analysis Plan	10%	How we'll interpret

Experiment Types

The Lean Startup profile supports various experiment types:

Type	When to Use	Effort
Smoke Test	Test demand before building	Low
Landing Page	Test value proposition	Low
Concierge	Manual service, learn process	Medium
Wizard of Oz	Fake automation, real experience	Medium
A/B Test	Compare specific variations	Medium
Prototype	Test usability, not demand	Medium
Cohort Analysis	Measure behavior over time	High

Example Workflow

# 1. Initialize project
multispec init food-delivery --profile lean-startup

# 2. Document hypotheses
multispec draft mrd -p food-delivery
# Focus on riskiest assumptions
multispec eval mrd -p food-delivery
multispec approve mrd -p food-delivery

# 3. Define MVP
multispec draft prd -p food-delivery
# Minimum features for learning
multispec eval prd -p food-delivery
multispec approve prd -p food-delivery

# 4. Design experiment
multispec draft uxd -p food-delivery
# How we'll test the hypothesis
multispec eval uxd -p food-delivery
multispec approve uxd -p food-delivery

# 5. Run experiment, collect data...

# 6. Document pivot/persevere
multispec draft narrative-1p -p food-delivery

# 7. If pivot, start new iteration
multispec init food-delivery-v2 --profile lean-startup

Metrics That Matter

Avoid Vanity Metrics

Vanity Metric	Actionable Alternative
Total users	Active users (DAU/MAU)
Page views	Conversion rate
Downloads	Retention rate
Registered users	Paying customers
Followers	Engagement rate

The One Metric That Matters (OMTM)

Choose one metric that best represents current learning focus:

• Problem validation: Customer interview conversion
• Solution validation: Sign-up rate
• Product validation: Activation rate
• Revenue validation: Paying conversion

Innovation Accounting

Track progress through three phases:

1. Establish baseline: First MVP data
2. Tune the engine: Improve toward ideal
3. Pivot or persevere: Decide based on data

Reference Materials

For deeper understanding of Lean Startup methodology, see:

• The Lean Startup
• The Lean Startup by Eric Ries
• Running Lean by Ash Maurya
• Internal reference: frameworks-internal/lean-startup/