Chatter velocity

The Chatter Velocity sub-tab — under the Engagement tab, next to Chatter Tiers — shows the shape of your stream: how many people were in chat at any moment, how quickly they arrived, and how they tailed off. It divides the session into 5-minute windows and reads everything from chat activity.

The Chatter Velocity sub-tab showing the three-series area chart and the per-bucket grid with tier breakdown bars
The Present area shows the room size at every moment; the running totals show when viewers arrived and departed.

Reading the chart

Three series are plotted across the stream:

SeriesWhat it shows
PresentUnique chatters active in each 5-minute window — the size of the room at that moment, drawn as a filled area
Arrived (total)A running total of unique chatters who have shown up so far. Where this line flattens is your total reach for the stream
Departed (total)A running total of chatters who have sent their last message — a cumulative drain curve

The peak window — when the most chatters were present at once — is marked with a vertical line and labelled with its time and count. Hover any point to see the present count and both running totals; the peak window is flagged with a ★.

How 'departed' is determined

A chat log records when someone last spoke, not when they stopped watching. The "Departed (total)" line therefore tracks last-message timing — silent lurkers who never chat again look like they departed at their last message, and by the end of the stream the line always climbs to the full unique-chatter count. Read it for the shape of the drain — when departures start to accelerate — rather than as an exact attendance figure.



For live sessions, the app uses an inactivity heuristic instead: a chatter is only counted as departed once they've been silent for 3 consecutive buckets (15 minutes). Chatters who last spoke within the last 2 buckets are still considered potentially present, so the curve won't climb during active stretches.

The per-window grid

Below the chart, one row per 5-minute window breaks the same data down:

ColumnMeaning
#Window number, in order
TimeThe elapsed-time range the window covers
ActiveUnique chatters present in the window
ArrivedChatters seen for the first time in this window
DepartedChatters whose last message falls in this window
NetArrivals minus departures for the window
PeakA ★ marks the peak window — the same one called out on the chart
Tier BreakdownA stacked bar showing the loyalty-tier mix (Core / Engaged / Casual / Drive-by) of the chatters active in that window

The Tier Breakdown bar uses the same four chatter tiers as the rest of the app, so you can see at a glance whether a busy window was carried by loyal regulars or a wave of one-off visitors.

Using it in practice

The Present curve answers "when was my stream at its biggest, and what shape was the session?" — a flat plateau, a single spike, or a slow build each tell a different story. A sharp jump in Arrived (total) often lines up with a raid or a clip going around. When the Departed (total) curve starts climbing steeply, that's where the room began to empty — worth comparing against what you were doing on stream at that time.

Pair it with highlights

Line up the peak window against your highlights — the moment chat was busiest is often, but not always, the moment it was loudest. When they differ, you've found a big-but-quiet stretch or a small-but-explosive one.