Activated
Crowd visibly reacting
Cheering, applause, dancing, bouncing, singing, clapping, hands up
The science
Show direction, crowd response, and fan conversion
Band.Coach measures observable behavior from live performance video and audio — not facial emotion scores or talent grades. Engagement per minute is our internal measure of show quality. Making more fans is the outcome bands care about.
Your songs are yours. We coach setlist order, visual spotlight, transitions, banter, movement, and look — never how to write your music.
Four audience states
Absorbed
Crowd quietly locked in
Stage orientation, coordinated stilling, low unrelated motion, sustained attention
Disconnected
Attention appears elsewhere
Conversation, backs turned, exits, long response decay, non-musical movement
Unknown
Camera can't tell
Occlusion, bad lighting, crowd out of frame
Engaged seconds per minute = measurable activated + absorbed seconds. Not “how much did the crowd move?” — and not a fixed percentage target.
A common performance-craft idea — leave room for the audience to participate — is not the same as a scientific “33% engagement” KPI. We learn the right response range for your band, genre, and room.
Fan outcomes — what bands care about
New-fan conversion
Did people who didn't know you follow, sign up, or listen afterward? (Requires post-show instrumentation — not camera alone.)
Post-show listen lift
Research on 1.3M listeners found concert attendance increased subsequent streaming of that artist.
Listen-again intent
Live presence correlates with intent to return and listen again — stronger than livestream alone.
Share / recommend intent
Immersion and satisfaction at live events link to loyalty and word-of-mouth.
Show mechanisms we track
Response occupancy
How much of the set showed observable participation or absorption — learned per band, genre, and room
Peak count & strength
Meaningful room-response moments relative to local baseline
Response latency
Seconds from performance cue → visible crowd response
Recovery time
Time to regain visible response after a drop
Repeat-show delta
Did experiments from last show produce measurable change?
What bands control (KAIs)
Set order & energy contour
Which song opens, peaks, resets, closes — without changing the songs themselves
Visual lead per section
Who owns audience attention during verse, chorus, solo, bridge
Song-to-song dead air
Seconds between applause/song end and next intentional moment
Banter duration & conversion
Which talking breaks produced response vs. which didn't
Spotlight on solos
Whether the band visually directs attention to the musical lead
Visual coherence
Focal hierarchy, contrast, lighting — not attractiveness scores
What we don't coach
- Stage presence: 76/100
- Move more / smile more
- Crowd engagement should be 33%
- Facial-emotion detection
- Telling bands their song is bad
- 20 recommendations after every gig
- Comparing movement across unrelated genres
Research we build on
Concerts increase subsequent listening
Natural experiment (1.3M listeners): attending a concert increased daily listening to that artist afterward, including non-fans becoming listeners.
Read more →Live vs. livestream immersion
2026 randomized experiment: physical presence produced greater immersion, listen-again intent, and physiological synchrony than livestream.
Read more →Movement is style-dependent
More body movement does not always equal better performance — optimal motion depends on musical style.
Read more →Visual attention & gaze
Eye-tracking shows performers can redirect audience attention; gaze toward listeners improves appreciation.
Read more →Familiarity in live settings
Field study (506 attendees): familiar, accessible pieces associated with stronger appreciation — relevant to set placement, not songwriting.
Read more →Audience motion & stillness
Live audiences show tempo-matched bouncing and dynamic stilling — silence can mean absorption.
Read more →
Band.Coach is a coaching product in development, not a peer-reviewed research instrument. Early pilots combine AI analysis with human review. We say “can't tell” when evidence is insufficient.