Manuscript Engagement Analytics

Core Lens

Reader engagement can be approximated by mapping value over reading time. A manuscript with long stretches between useful payoffs, vague topic headings, or reader-comment dropoff is signaling where readers may get bored, confused, or stuck.

Use this skill to:

• Generate heading-level word-count maps.
• Find slow starts and long slogs.
• Audit whether headings promise reader takeaways.
• Interpret beta-reader comment locations and abandonment.
• Produce a revision queue for value pacing.

Reference Routing

Script

Use scripts/analyze_manuscript.py for deterministic Markdown structure analysis:

python3 skills/manuscript-engagement-analytics/scripts/analyze_manuscript.py manuscript.md

It outputs a table of headings, line numbers, word counts, cumulative words, and heuristic flags. Use the script output as evidence, then apply judgment from the references.

Workflow

1. Establish The Reader Promise

Identify the target reader, book promise, and first meaningful payoff. If these are unclear, use book-toc-lab first.

2. Generate A Structure Map

Run the script or manually build a table:

Section | Line | Words | Cumulative words | Reader takeaway | Risk

3. Mark Value Events

Mark where the reader gets:

• A usable idea.
• A decision frame.
• A checklist.
• A worked example.
• A lab or exercise.
• A troubleshooting answer.

4. Diagnose Engagement Risks

Look for:

• Too many words before first payoff.
• Long sections with weak takeaways.
• Back-to-back setup sections.
• Vague headings that hide the reader value.
• Beta-reader comments stopping near the same section.

5. Recommend Revision Actions

Prefer structural fixes:

• Move value earlier.
• Cut or compress low-payoff setup.
• Rename headings around reader outcomes.
• Split long sections.
• Convert theory into examples, checklists, labs, or decisions.

Output Format

When auditing engagement, return:

1. Promise and first-payoff diagnosis.
2. Value map or script output summary.
3. Highest-risk sections.
4. Revision recommendations ordered by expected engagement impact.
5. Beta-reader comment/dropoff interpretation when data exists.
6. Follow-up checks after revision.

Quality Bar

Use metrics as signals, not verdicts. Word counts and comment dropoff show where to inspect; the final recommendation should explain what reader value is missing, delayed, or unclear.