Lean product development minimizes waste by validating hypotheses before fully building. Instead of months of development, launch quickly, learn and iterate. This guide presents lean development principles and practices.
Origins of Lean
Toyota Production System
Lean manufacturing. Waste elimination.
Lean Startup
Eric Ries applied to the software. Build-Measure-Learn.
Core Principles
Validation, iteration, continuous learning.
Build-Measure-Learn
Build
Build the minimum necessary to test hypothesis.
###Measure
Collect data on actual usage.
Learn
Insights for next iteration.
###Loop
Continuous and fast cycle.
MVP (Minimum Viable Product)
Definition
Minimum version that allows validated learning.
Objective
Test hypotheses, not impress.
Examples
Landing page, prototype, wizard of oz.
Traps
Too minimal or too viable.
Validation of Hypotheses
Value Hypothesis
Does it solve a real problem?
Growth Hypothesis
Can you acquire users?
Experiments
Test before committing.
Metrics
How will you know if it has been validated?
Types of MVP
Landing Page
Value proposition + interest capture.
Fake Door
Button that measures interest before building.
Concierge
Do manually what you would automate.
###Wizard of Oz
It looks automatic, it's manual behind it.
Single Feature
A well done feature.
Piecemeal
Use existing tools to simulate.
Metrics that Matter
Vanity Metrics
Downloads, pageviews. They look good, they don't indicate success.
Actionable Metrics
Activation, retention, revenue. They guide decisions.
North Star
Single metric that represents value.
Leading Indicators
Predictors of outcome metrics.
Pivots
What is it
Strategic change based on learning.
Types
Segment, channel, technology, feature.
When to Pivot
Core hypotheses repeatedly invalidated.
Perseverance vs Pivot
Difficult decision. Data helps.
Continuous Discovery
Integration
Continuous research, not an isolated phase.
Dual Track
Discovery and delivery in parallel.
Weekly Habits
Search every week.
Prototyping
Test before code.
Lean Canvas
What is it
Simplified business model canvas.
Sections
Problem, solution, metrics, unique value proposition.
Usage
Document and iterate business model.
Reducing Waste
Code Nobody Uses
Features that do not generate value.
Time in Meetings
Excessive alignment.
Analysis without Action
Data paralysis.
Premature Perfection
Polish before validating.
Kanban for Product
WIP Limits
Limit work in progress.
###Flow
Optimize flow, not utilization.
Pull
Pull work when there is capacity.
Preview
Make work visible.
Experiments
Hypothesis
"We believe [change] will cause [outcome] because [reason]."
Design
How to test? What to measure?
Execution
Run the experiment.
Analysis
What have we learned?
Team Lean
Small
Less people, more speed.
Multifunctional
All necessary skills.
Autonomous
Make decisions without needing approval.
Focused
One goal at a time.
Lean Deliverables
Minimum Documentation
Enough to communicate.
Just-in-Time Design
Don't draw everything in advance.
Iterative Code
Refactor as you learn.
Common Errors
MVP as an Excuse
For bad product. MVP must be usable.
No Metrics
Launch and don't measure.
Lots of Build, Little Learn
Too long cycles.
Ignore Qualitative
Numbers without understanding why.
Conclusion
Lean product development is a learning discipline. Minimize time between idea and feedback, validate before building too much and be willing to change. The result is products that really solve problems.
##FAQs
1) Does Lean work for large companies? Yes. Principles apply. Adapted implementation.
2) MVP means bad product? No. Minimal but functional and usable.
3) How do you know when to pivot? Data repeatedly invalidating central hypotheses.
4) Is Lean the same as Agile? Related but different. Lean is validation, Agile is execution.
5) How long should a cycle last? Weeks, not months. As quickly as possible.
Also read
- Product Validation: How to Test Ideas Before Building
- Lean Product Development: Planning for Startups
- Lean product development for beginners: plan to learn, not to get it right the first time
- Zero to One: How to Launch a Digital Product from Zero
- Data Driven Product
- Events and Tracking: Analytics for Digital Products