🔬 Hypothesis Generation
Guide to scientific hypothesis generation and test methodology.
📋 Hypothesis Structure
Format
IF [independent variable/action]
THEN [dependent variable/outcome]
BECAUSE [mechanism/reasoning]
Example
IF we reduce checkout steps from 5 to 3
THEN conversion rate will increase by 15%
BECAUSE fewer steps reduce friction and drop-off
🎯 Hypothesis Criteria
| Criterion |
Description |
| Specific |
Clear and no ambiguity |
| Measurable |
Measurable outcome |
| Testable |
Testable |
| Falsifiable |
Falsifiable |
| Relevant |
Aligned with business goal |
🔧 Hypothesis Types
A/B Test Hypothesis
**Hypothesis:** Changing CTA button from blue to green
will increase click rate by 10%
**Metric:** CTA Click Rate
**Baseline:** 2.5%
**Target:** 2.75%
**Sample Size:** 10,000 users
**Duration:** 2 weeks
Product Hypothesis
**Problem:** Users abandon during onboarding
**Hypothesis:** Adding progress indicator will reduce
abandonment by 20%
**Success Metric:** Onboarding completion rate
📊 Experiment Design
Test Plan
## Experiment: [Name]
### Hypothesis
[IF-THEN-BECAUSE statement]
### Variables
- Independent: [What we change]
- Dependent: [What we measure]
- Control: [What stays same]
### Metrics
- Primary: [Main KPI]
- Secondary: [Supporting metrics]
- Guardrail: [Safety metrics]
### Design
- Type: A/B / Multivariate
- Split: 50/50
- Duration: [X] weeks
### Analysis Plan
- Statistical test: [t-test, chi-square, etc.]
- Confidence level: 95%
- MDE: [Minimum detectable effect]
📝 Prioritization (ICE)
| Hypothesis |
Impact |
Confidence |
Ease |
Score |
| H1 |
8 |
7 |
6 |
7.0 |
| H2 |
9 |
5 |
4 |
6.0 |
| H3 |
6 |
8 |
9 |
7.7 |
ICE Score = (Impact + Confidence + Ease) / 3
Hypothesis Generation v1.1 - Enhanced
🔄 Workflow
Source: Stanford d.school Design Thinking
Stage 1: Observation
- Data: Capture an "Insight" from analytics data or a user interview.
- Problem: Turn the observation into a clear problem statement.
Stage 2: Construction
- Formula: Use the template IF [action] THEN [outcome] BECAUSE [reason].
- Variables: Clarify the independent (changing) and dependent (measured) variables.
Stage 3: Prioritization
- ICE Score: Score Impact, Confidence, and Ease from 1-10.
- Risk: What do we lose if the test fails?
Checkpoints
| Stage |
Validation |
| 1 |
Can the hypothesis be falsified? (If it is always true, it is not a hypothesis) |
| 2 |
Is the outcome a measurable metric (Click rate, Retention)? |
| 3 |
Is the "Because" part based on logical user behavior? |
