Analytics
Read the PanelWave analytics dashboard — reads, completion, reading time, audience breakdowns, the reading funnel, per-panel engagement, the click heatmap, and CSV export.
The Analytics Dashboard shows how readers engage with a published work: how often it is read, how far readers get, and how long they stay. Open a work's backstage and choose Analytics in the navigation (it is also available from the work card's More actions menu in the works list).
Metrics are aggregated nightly from player reading events. A work only starts collecting data once a published version is being read — a brand-new or unpublished work shows "No reader data yet".
Choosing what you look at
The header reads "Reader metrics for this work — last 30 days" (or whichever range is active). Controls at the top let you slice the data:
- Time range — preset buttons for 7 days, 30 days (default), or 90 days, plus Custom with start/end date pickers.
- Compare with previous period — a checkbox that fetches the preceding window of equal length and adds change badges (e.g. +25% / −12%) to every metric; metrics that had no data before show a new badge.
- Device — filter to a single device type, or All devices.
- Locale — filter to a single reading language, or All locales.
Two buttons complete the header: Refresh and Export CSV.
The four key metrics
Each metric is a tile with the headline number, a small trend sparkline for the selected range, and a supporting figure:
| Tile | What it means |
|---|---|
| Total reads | Number of reading sessions in the range; the sub-line shows how many were from signed-in readers |
| Avg completion | Average share of the work a session got through (e.g. 83%); the sub-line shows average panels per session |
| Total reading time | Combined time readers spent in the work (e.g. "2h 5m") across the active days in the range |
| Avg session | Average length of a single reading session |
Per-day detail
Click any tile to open the "Per-day detail" table for the current range and filters, with columns Date, Reads, Signed-in, Completion, Reading time, Avg session, and Panels/session — the column for the metric you clicked is highlighted. Click Close to collapse it. If the selected slice has no data the table says "No sessions in this range."
Audience breakdowns
Below the tiles, two panels show where your sessions come from:
- Devices — sessions per device type, as horizontal bars with counts.
- Locales — sessions per reading language.
Geographic (country-level) data is not collected — device and locale are the two audience dimensions available.
Reading funnel
The Reading funnel shows where readers drop off, panel by panel in reading order: each step is a panel with the number of sessions that reached it, and the header summarizes how many sessions completed the work (e.g. "12 of 20 sessions completed (60%)"). A steep drop between two steps is your clearest signal for a pacing problem, a confusing branch — or a well-placed paywall boundary.
Panel engagement
Panel engagement ranks your panels by views, with a bar chart on top and a table below showing per-panel Views, Avg dwell (how long readers stay on the panel), and Bounce rate (sessions that saw only this panel and left). Long dwell can mean a beloved splash page — or a panel readers get stuck on; pair it with the funnel to tell the difference.
Funnel and per-panel metrics come from the same nightly aggregation as the key metrics — they appear once reading sessions have been processed.
Click heatmap
The Click heatmap shows where readers click on a panel — every tap, whether it landed on a hotspot or not:
- Pick a panel from the Panel selector (it lists the panels that have reader activity).
- The panel renders with a density overlay: cool blue where a few readers clicked, hot red where many did. Defined hotspot regions are outlined with their click counts.
- The header summarizes the split — e.g. "120 clicks — 85 on hotspots, 35 dead (29%)" — and the Hotspot hits list ranks your hotspots by clicks.
Dead clicks (taps that hit no hotspot) are the interesting part: a hot cluster outside every hotspot means readers expect something there to be interactive — consider putting a hotspot where they're already tapping, or making the existing one bigger. A hotspot with few hits despite traffic may be too subtle or oddly placed.
Unlike the other sections, the click heatmap reads raw click events, so fresh clicks show up without waiting for the nightly aggregation. Very busy panels are sampled to the most recent 5,000 clicks (the header says "sampled" when that happens).
Exporting the data
Export CSV downloads the per-day detail for the current range and filters as a CSV file (one row per day, plus a totals row) — handy for your own spreadsheets or sharing with collaborators.
Interpreting the numbers
- A read = a session. One reader returning three times counts as three reads; signed-in readers is the closest thing to unique people.
- Completion tells you whether readers finish. A high read count with low completion usually means readers drop off — check pacing, chapter length, or where your paywall preview boundary sits.
- Panels per session vs. your total panel count gives a rough sense of how deep a typical session goes.
- Compare with previous period is the quickest way to judge whether a new chapter, a promotion, or a price change moved the needle — look for the green/red change badges.
- Filtering by device can reveal layout problems: if phone sessions complete far less than desktop sessions, preview the work in the phone frame and check panel legibility.
- Filtering by locale shows whether a translation is pulling its weight.
- The click heatmap is your interactivity audit: dead-click clusters show where readers want a hotspot; a cold hotspot is one they can't find.
Because aggregation runs nightly, don't expect today's reads to appear immediately — check back the next day when judging a release or announcement.
Data retention — raw reading events are automatically deleted after 24 months; the aggregated per-day metrics that power this dashboard are kept, so long-range trends survive the cleanup.