Survey Analysis Framework (AI Prompt)

## Survey Analysis Framework

**Purpose:** Understand churn after 3 months
**Responses:** 347
**Decision:** Onboarding vs. product features investment

---

### ANALYSIS PROCESS

**Step 1: Data Cleaning (30 min)**
□ Remove incomplete responses (<50% complete)
□ Check for obvious spam/gibberish
□ Normalize text (lowercase for analysis)
□ Flag outliers for review

**Step 2: Overview Analysis (1 hour)**
□ Calculate overall NPS score
□ Average ratings for each question
□ Response distribution for multiple choice
□ Completion rate by question

**Step 3: Segmentation Analysis (2 hours)**
□ Split by tenure (0-1 mo, 1-3 mo, 3+ mo)
□ Split by plan type (if applicable)
□ Split by use case / persona
□ Compare segments on key metrics

**Step 4: Deep Dive Analysis (2-3 hours)**
□ Cross-tabulate key questions
□ Identify statistically significant differences
□ Code open-ended responses
□ Extract representative quotes

---

### QUANTITATIVE ANALYSIS

**NPS Analysis:**
| Segment | Promoters | Passives | Detractors | NPS |
|---------|-----------|----------|------------|-----|
| Overall | ? | ? | ? | ? |
| 0-1 mo | ? | ? | ? | ? |
| 1-3 mo | ? | ? | ? | ? |
| 3+ mo | ? | ? | ? | ? |

*Look for: Where does NPS drop significantly?*

**Statistical Significance:**
- With 347 responses, differences >5% are likely significant
- For segment comparisons, need n>30 per segment
- Use chi-square for categorical, t-test for numeric

**Correlation Mapping:**
Create a matrix: Which factors correlate with satisfaction?
- Onboarding experience vs. NPS
- Feature usage vs. satisfaction
- Support interactions vs. churn intent

---

### QUALITATIVE ANALYSIS

**Open Response Coding Framework:**

1. Read all responses once without coding
2. Create initial code list based on themes
3. Code each response (can have multiple codes)
4. Group codes into themes
5. Count frequency of each theme
6. Select representative quotes (1 per theme)

**Sample Code List for Churn Survey:**
| Code | Theme | Frequency |
|------|-------|-----------|
| ONBOARD-CONFUSE | Onboarding Issues | ? |
| ONBOARD-TIME | Onboarding Issues | ? |
| FEATURE-MISSING | Product Gaps | ? |
| FEATURE-COMPLEX | Product Issues | ? |
| SUPPORT-SLOW | Support Issues | ? |
| PRICE-HIGH | Pricing | ? |
| COMPETITOR | Competition | ? |

---

### INSIGHT EXTRACTION TEMPLATE

**From Data to Insight:**

*Observation:* [What the data shows]
*Insight:* [What it means]
*So What:* [Why it matters]
*Recommendation:* [What to do about it]

**Example:**
*Observation:* NPS drops from 45 to 12 between month 1 and month 3
*Insight:* Customers are happy initially but lose enthusiasm as they try to use advanced features
*So What:* We're losing customers at the critical "activation" phase
*Recommendation:* Create an onboarding sequence specifically for advanced features at week 4...