# Hypothesis Generation

> Bilimsel hipotez oluşturma, deney tasarımı ve test metodolojisi rehberi.

- Skill: `majiayu000/hypothesis-generation` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add majiayu000/hypothesis-generation`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/hypothesis-generation/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools, AI & ML
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/hypothesis-generation

---


# 🔬 Hypothesis Generation

> Bilimsel hipotez oluşturma ve test metodolojisi rehberi.

---

## 📋 Hipotez Yapısı

### Format
```
IF [independent variable/action]
THEN [dependent variable/outcome]
BECAUSE [mechanism/reasoning]
```

### Örnek
```
IF we reduce checkout steps from 5 to 3
THEN conversion rate will increase by 15%
BECAUSE fewer steps reduce friction and drop-off
```

---

## 🎯 Hipotez Kriterleri

| Kriter | Açıklama |
|--------|----------|
| **Specific** | Net ve belirsizlik yok |
| **Measurable** | Ölçülebilir outcome |
| **Testable** | Test edilebilir |
| **Falsifiable** | Yanlışlanabilir |
| **Relevant** | İş hedefine uygun |

---

## 🔧 Hipotez Türleri

### A/B Test Hipotezi
```markdown
**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 Hipotezi
```markdown
**Problem:** Users abandon during onboarding
**Hypothesis:** Adding progress indicator will reduce 
abandonment by 20%
**Success Metric:** Onboarding completion rate
```

---

## 📊 Experiment Design

### Test Plan
```markdown
## 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

> **Kaynak:** [Stanford d.school Design Thinking](https://dschool.stanford.edu/resources)

### Aşama 1: Observation
- [ ] **Data**: Analitik verisi veya kullanıcı görüşmesinden bir "Insight" yakala.
- [ ] **Problem**: Gözlemi net bir problem cümlesine dönüştür.

### Aşama 2: Construction
- [ ] **Formula**: IF [action] THEN [outcome] BECAUSE [reason] şablonunu kullan.
- [ ] **Variables**: Bağımsız (değişen) ve bağımlı (ölçülen) değişkenleri netleştir.

### Aşama 3: Prioritization
- [ ] **ICE Score**: Impact (Etki), Confidence (Güven), Ease (Kolaylık) 1-10 puanla.
- [ ] **Risk**: Test başarısız olursa ne kaybederiz?

### Kontrol Noktaları
| Aşama | Doğrulama |
|-------|-----------|
| 1 | Hipotez yanlışlanabilir mi? (Her zaman doğruysa hipotez değildir) |
| 2 | Sonuç ölçülebilir bir metrik mi (Click rate, Retention)? |
| 3 | "Because" kısmı mantıklı bir kullanıcı davranışına dayanıyor mu? |