Drop-off Analytics v2: Spot the chapters where readers truly quit
Not every dip means abandonment. V2 clears out noise and focuses on true drop-off points:
- Right-censors the final chapter: avoids mistaking “end of series” as a mass drop-off.
- Maturity window (≥7 days, configurable): counts drop-off at chapter i only if i+1 has been public long enough.
- Observed publish time: inferred from first reader activity, no DB schema changes required.
- Smarter peak detection: skips censored points while preserving smoothing logic.
- Developer-friendly: old APIs stay compatible, with optional new params and helper
init_manga_dropoff_delete_cache().
Impact: Identify the exact chapter that breaks retention, fix pacing or content issues, and recover reader loyalty faster.
Now Reading (Live Pulse): Minute-by-minute heartbeat of your site
A new ranking tab for admins showing what’s being read right now:
- Minute-decay scoring: points fade over time, capturing current heat.
- Per-manga live viewer count: not just views — shows how many people are actively reading each title this minute.
- Always-on fallback: if traffic is low, auto-switches to “recent chapters” or “24h uniques” to ensure results never go empty.
- Lightweight: lazy-init (loads only when clicked), UIKit dot indicator, and 45s transient cache to minimize query load.
Quick win: instantly spot which manga are “alive” right now, promote them on banners, or measure immediate reaction after a release.
Genre Engagement Index: Which genres truly hold attention?
Popular doesn’t always mean sticky. This dashboard widget (admin-only) creates a compact yet powerful engagement score per genre:
- Unique readers per genre + reading depth (avg chapters per reader).
- Mini-retention: percentage of readers hitting ≥3 or ≥10 chapters in the same genre.
- Engagement Index: combined formula (unique × depth × retention) for cross-genre comparison.
- Smart display: shows only top genres (default 12), cached for 15 minutes, and only registers widget if user has permission.
Quick win: double down on genres with high stickiness, adjust homepage banners, or plan release calendars by genre appetite.
3-Step Playbook (weekly growth routine)
- Check Live Pulse 15–60 minutes after new chapter drops to detect what’s heating up.
- Audit Drop-off v2 on series with declining numbers, fix pacing/cliffhangers at the true exit chapter.
- Use Genre Index to pick 2–3 “sticky” genres and align homepage/marketing pushes accordingly.
Performance & operations
- No schema changes, reuses
manga_reading_historytable. - Caching: Live Pulse 45s · Genre Index 15m · Drop-off cached per series/param set.
- Lean queries: grouped by time window, top-only selection, no heavy cron jobs.
Conclusion
v1.5.0 is about seeing the heartbeat of your readers in real time, spotting the true drop-off points, and understanding which genres drive engagement. Less guessing, faster response, and more sustainable growth for your manga community.
Comments