Lean Startup Notes

Hypothesis-Driven Entrepreneurship: The Lean Startup

Startups use hypothesis-driven entrepreneurship to handle resource limits and uncertainty. This method turns a vision into testable hypotheses and uses Minimum Viable Products (MVPs) to test them. Based on test results, the business model is either kept, changed, or dropped. This goes on until the business model is proven by MVP tests. After this, the startup has reached product-market fit and can start scaling.

Hypothesis-Driven Approach

This approach lowers the risk of making a product no one wants. Lean startups focus on learning how to build a lasting business instead of quick growth. Eric Ries named lean startup to describe orgs using hypothesis-driven entrepreneurship, avoiding waste which optimizes use of resources.

Lean Startup Method

Lean startup accelerates innovation using rapid iteration and short cycle times. The process includes:

  1. Creating hypotheses.

  2. Testing the hypotheses.

  3. Acting on test feedback.

Vision Development

Before making business model hypotheses, have a vision for the startup's problem and possible solution. Appendix B has vision generating guidlines.

Hypotheses Translation

Turn the vision into testable business model hypotheses. A business model includes choices about customer value, activity setup, and profit earning.

Falsifiability

Make falsifiable hypotheses for each part of the business model. A hypothesis can be rejected with an experiment. Without a falsifiable hypothesis, learning isn't validated.

Quantitative Metrics

Use quantitative metrics to validate hypotheses when possible. Monitor the customer conversion funnel:

  • The process through which a prospect becomes a customer.

  • Combine funnel data to estimate customer lifetime value (LTV) minus acquisition cost (CAC).

  • Use cohort analysis to track LTV/CAC trends.

A/B Testing

Use A/B tests. A/B tests divide similar prospects into a control group (status quo product) and a treatment group (product with a change). A/B testing sees if changes improve performance.

Comprehensiveness

At first, detailed hypotheses are not needed for all parts of the business model because of the serial dependence between business model elements. To avoid this, quickly check all parts of the business model to find possible problems early.

MVP Tests

To deal with uncertainty and limited resources, maximize learning per unit of time. Investor Paul Graham says to "launch early and often".

Minimum Viable Product (MVP)

An MVP has the fewest features and activities needed to test a business model hypothesis. MVPs reduce product development batch sizes and cycle times. This leads to fast feedback and easier problem diagnosis.

MVP Constraints

MVPs may limit: product functionality and/or operational capability. Minimal MVPs can be smoke tests that limit both functionality and operations.

Constrained Functionality

Limit MVP product functionality when:

  • Early adopters will buy a product with "need to have" but not "nice to have" features.

  • Some early adopters (Group A) won't use certain costly features deemed "need to have" by other early adopters (Group B).

Constrained Operations

The technology delivering the MVP's function is often temporary. Use MVPs with limited operational capability when defining the product's core function.

Series of MVPs

Entrepreneurs can validate key business model assumptions iteratively. Rent the Runway (RTR) used a series of MVPs.

Constraining Customer Sets

Test MVPs with a smaller customer set. Acquiring many customers before validating hypotheses can be costly and damage the brand.

False Positive and Negative Results

When making MVPs, consider these risks:

  • False positive: Indicates the hypothesis is confirmed, but it is not valid.

  • False negative: Indicates the hypothesis is disconfirmed, but it is valid.

Concerns about MVPs

Concerns include:

  • Idea theft: Early feedback outweighs the risk of competition. Ideas are worthless without execution.

  • Reputational risk: MVP testing limits the customer base to mitigate quality problems. Different brand names can also lower risks.

Test Prioritization

Prioritize tests eliminating risk at a low cost. When elements are serially dependent, sequencing experiments is important. Parallel testing involves tradeoffs. If a hypothesis is rejected, effort is wasted. However it can confer benefits in winner take all markets. When a startup faces threat of possible preemption, it helps gain a time to market edge by parallel testing.

Learning From MVP Tests

Entrepreneurs can evaluate feedback from MVP tests. Be aware of:

  • False positives or negatives.

  • Customer's stated vs true preferences.

  • Potential errors in interpreting feedback and cognitive biases.

Unexpected test results

Customers frequently behave in unanticipated ways when using a new product.

Information not derived from testing

Revise business model hypotheses using sources beyond test results, such as:

  • Competitor's announcements.

  • Regulator's actions.

  • News about new technologies.

Entrepreneurs should note surprises because mechanically monitoring data may cause missed opportunities or unforseen threats.

Persevere, Pivot, or Perish

After MVP tests, decide to:

  • Persevere: If the MVP validates the hypothesis and feedback doesn't prompt a change, continue.

  • Pivot: If the MVP test rejects the hypothesis, but feedback indicates opportunity, change some business model elements while keeping others.

  • Perish: If a MVP test rejects a key hypothesis and pivoting is not an option, shut down the business.

Scaling and Ongoing Optimization

If the entrepreneur has validated all key hypotheses, achieving product-market fit, scale and invest greatly in customer acquisition. Continue tests for business model optimization.

Limits to Lean Startup Methods

The hypothesis-driven approach is applicable to many new ventures. There are some situations where the lean startup process yields fewer advantages:

  • When mistakes must be limited.

  • When customer demand is low.

  • When product development cycles preclude launching early and often

Alternative Approaches for Launching Startups

  • Build It And They Will Come: Focus on product development without customer feedback; risky if demand is uncertain.

  • Waterfall Planning: Sequential stages with formal reviews; rigid and slow to adapt to change.

  • Just Do It!: Improvisational approach, adapts to feedback but lacks initial direction; can be inefficient.

Key Terms

  • A/B testing: Divides prospects into control and treatment groups to test modifications.

  • Business model: Choices defining customer value, operations, go-to-market, and cash flow.

  • Cohort analysis: Tracks performance metrics using customer groups acquired over time.

  • Conversion funnel: Process through which prospects become loyal customers.

  • False negative: Test incorrectly indicates hypothesis is invalid.

  • False positive: Test fails to disprove invalid hypothesis.

  • Falsifiable hypothesis: Can be disconfirmed through experiment.

  • Lean startup: Tests business model hypotheses using MVPs to avoid waste.

  • Minimum Viable Product: Fewest activities needed to disprove hypothesis.

  • Pivot: Changing some business model elements.

  • Product-market fit: Right product for the market with demand and profit potential.

  • Smoke test: Gauges demand for a non-existent product.