# Reader Feedback Analysis & Creative Decision Loop

## I. Feedback Collection Channels

| Channel | Content Collected | Frequency |
|---------|-------------------|-----------|
| Chapter comments / inline comments | Instant per-chapter reactions | Daily |
| Comment section | Overall evaluation, plot discussion | Daily |
| Reader community / group | Deep discussion, character preferences | Weekly |
| Backend data | Retention rate, stickiness, reading duration | Weekly |
| Reviews / ratings | Structured evaluation | Monthly |

## II. Feedback Classification & Weighting

### By Credibility

```
S-level — Backend data (retention rate, stickiness): objective, indisputable
A-level — High-frequency repeated feedback (≥5 people mention the same issue): likely real problem
B-level — Specific criticism with logical basis: worth serious consideration
C-level — Emotional venting without specifics: low reference value
D-level — Demands contradicting core setting: reject on principle
```

### By Impact Scope

```
Systemic issues — Affect majority of readers → priority handling
Local issues — Affect some readers/single chapter → handle as appropriate
Personal preference — Individual reader subjective feeling → reference only
```

## III. Feedback Analysis Template

```
【Analysis time】 After Ch. X publication
【Data summary】
- 24h retention rate: XX% (vs. previous chapter: ±X%)
- Chapter comments: XX (positive/neutral/negative = X/X/X)
- Core discussion topics (TOP3):
  1. XXX
  2. XXX
  3. XXX

【Key feedback】
- Positive (why readers like):
- Negative (what readers are unhappy about):

【Problem diagnosis】
- Quality issue? (writing/pacing/logic)
- Content direction issue? (plot direction/character fate doesn't match expectations)
- Update issue? (breaks/less frequent updates/slow pace)

【Creative decision】
Adopt / Partially adopt / Reject
Reasoning:
Action plan:
```

## IV. Decision Principles

### Should Listen To

- Retention rate dropping → must find cause and adjust
- ≥5 people mention the same specific problem → likely real issue
- "Don't understand" "can't follow" → information delivery problem
- "Too padded" "too slow" → pacing needs adjustment

### Should NOT Blindly Follow

- Demands to change core setting (shakes the book's foundation)
- Demands "write more of character X" when X shouldn't appear yet in plot
- "Hurry up and get together / defeat the villain" → pacing plan shouldn't be disrupted
- Extreme emotional attacks → don't respond, don't internalize

### Flexible Handling

- Dialogue style / description density adjustment (doesn't affect core setting)
- Minor adjustment to supporting character screen time (without affecting main plot)
- Update frequency fine-tuning (within capability)

## V. Decision Log & Tracking

```
═══════════════════════════════════════
【Creative Decision Log】

Decision #01  |  Date: XXXX-XX-XX
Source: Ch. X comments (retention dropped from 75% to 68%)
Issue: Readers report "too many daily chapters, main plot advancing too slowly"
Decision: Compress daily plotline over next 5 chapters, accelerate main plot
Execution:
  - Ch. X: Cut daily dialogue XXX
  - Ch. Y: Compress 3-chapter setup into 2 chapters
  - Ch. Z: Bring forward main plot conflict introduction
Verification: After adjustment, retention rate recovered to 73%
═══════════════════════════════════════
```

## VI. Feedback Loop Check

After each adjustment, observe at least 5 chapters, confirm:

- [ ] Did data improve (retention/stickiness)?
- [ ] Did negative feedback decrease?
- [ ] Did adjustment introduce new problems?
- [ ] Does it deviate from outline's main line?

If adjustment ineffective → possibly misdiagnosed the problem, re-analyze.

## VII. Mindset Management

- **Not all criticism needs response**: web novel reader base is large, can never satisfy everyone
- **Data > Intuition > Individual comments**: one person's problem may just be personal preference
- **Writing well > Writing fast**: rushing to win back readers often backfires
- **Logging beats remembering**: decision log is the most valuable long-term creative asset