Use this dashboard to understand how live video sessions perform—who joins, how long they stay, when they churn, and where quality or engagement drops create retention risk. Pair with moderation and raw export tools for deeper investigation.
Live Streams
Volume & scheduling
Audience Reach
Unique viewers
Watch Time
Minutes & avg durations
Peak Concurrency
Capacity & scaling
Engagement
Chat & reactions
Quality
Bitrate / resolution distribution
Retention
Viewer stickiness
Conversion / CTA
Down‑funnel actions*
Safety
Flagged segments/messages
* Conversion metrics appear only if CTAs or monetization events are enabled.
Fast Questions You Can Answer
Are we producing enough streams to meet audience demand?
Daily Operational Workflow
1
Select Scope
Pick date range & segment (region, category, creator cohort).
2
Scan KPI Cards
Streams, Unique Viewers, Minutes Watched, Peak Concurrency deltas vs prior period.
3
Check Concurrency Curve
Look for sharper peaks or plateau changes requiring scaling adjustments.
4
Evaluate Retention Milestones
Inspect watch duration distribution to spot early drop-off points.
5
Review Engagement Rate
Compare chat messages & reactions per viewer to baseline.
6
Quality Distribution
Monitor resolution tiers (FHD/HD/SD) & bitrate shifts; investigate abrupt downgrades.
7
Top Streams Leaderboard
Identify exemplary or underperforming sessions—note anomalies.
8
Moderation & Safety
Check flagged content count & ratio; queue backlog if present.
9
Log Interventions
Record actions (promotion, creator coaching, infra tuning) and schedule follow-up.
Modules & Interpretation
KPI Cards
Provide macro health: production (Live Streams), audience (Viewers), consumption (Minutes Watched), engagement (Reactions / Chat), and performance (Peak Concurrency).Concurrency Trend
Visualizes simultaneous viewers across sessions/time. Rising volatility without capacity scaling → risk of buffering or join failures.Watch Time Distribution
Shows tails of viewing session durations. Healthy distributions shift gradually right as content relevance improves.Top Live Streams Table
Leaderboard enabling drill-down into Stream Title, Host, Start Time, Peak Concurrent Viewers, Unique Viewers, Minutes Watched, Engagement Rate, Flagged Items.Resolution / Quality Mix
Breakdown of minutes by FHD, HD, SD. Persistent drift toward SD may reflect network constraints or encoding fallback issues.Engagement Over Time
Overlay of chat messages & reactions against viewer count to spot interaction troughs where hosts can re-engage.Safety & Moderation Panel
Flagged chat messages / segments, automated policy triggers, and unresolved moderation queue aging.Key Metrics & Formulas
| Metric | Definition | Formula (Illustrative) | Why It Matters | Action Trigger |
|---|---|---|---|---|
| Live Streams | Unique live broadcast sessions | count(stream_id) | Supply of content | Drop >20% w/w |
| Unique Viewers | Distinct users viewing ≥1s | distinct(user_id) | Reach & growth | Stagnant vs new user growth |
| Minutes Watched | Total minutes across all viewers | Σ session_minutes | Engagement depth | Decline without stream count drop |
| Avg Minutes per Stream | Consumption per broadcast | Minutes Watched / Live Streams | Stream effectiveness | < baseline -15% |
| Avg Minutes per Viewer | Viewer stickiness | Minutes Watched / Unique Viewers | Retention quality | Early drop-off spike |
| Peak Concurrent Viewers (PCV) | Max simultaneous viewers | max(concurrent_viewers_t) | Capacity planning | >80% infra headroom |
| Concurrency Growth Rate | % change in PCV vs prior period | (PCV_now - PCV_prev)/PCV_prev | Scaling forecast | >25% sustained increase |
| Chat Messages | Total chat messages in streams | count(chat_msg) | Interactive engagement | Flat while viewers rise |
| Reactions | Total live reactions | count(reaction) | Lightweight engagement | Ratio down vs baseline |
| Engagement Rate | (Chat + Reactions)/Unique Viewers | (C + R)/Viewers | Content resonance | Falls below target band |
| Flagged Items | Flagged chat or segments | count(flags) | Safety risk | Spike >30% w/w |
| Flagged Rate | Flagged / Chat Messages | flags / chat_msg | Moderation workload | > threshold (e.g., 2%) |
| FHD % | Share of minutes at FHD | FHD_minutes / total_minutes | Quality experience | Sudden drop >10 pts |
| Rebuffer Complaint Rate* | User reports / viewer sessions | complaints / sessions | QoE stability | Rise > baseline +2σ |
| Start Failure Rate* | Failed stream starts / attempts | fails / attempts | Onboarding friction | >1–2% sustained |
| Retained 5m Viewers | Viewers staying ≥5 minutes | viewers_5m / Unique Viewers | Early retention | Drop >10 pts |
| Retained 25% Duration* | Viewers watching ≥25% of stream length | viewers_25pct / Unique Viewers | Session stickiness | Falling trend |
* Metrics appear only if enabled in your telemetry configuration.
Analytical Lenses
Unique viewers & new viewer share.
Benchmark & Governance Strategy
Baseline Period
Baseline Period
Collect first stable 4 weeks post-launch to set medians & IQR.
Adaptive Alerts
Adaptive Alerts
Trigger alert if metric leaves median ±1.5 * IQR (robust to outliers).
Retention Milestones
Retention Milestones
Track % reaching 1m, 5m, 10m; focus improvements where steepest drop exists.
Quality SLAs
Quality SLAs
Define minimum FHD% or acceptable SD% band; escalate on rapid degradation.
Capacity Planning
Capacity Planning
Forecast PCV at 4–6 week horizon using rolling CAGR to schedule scaling windows.
Safety Policy
Safety Policy
Escalate if flagged rate breaches threshold twice within 7 days.
Early Warning Signals
High viewers; low chat & reactions → prompt host engagement tactics.
Troubleshooting
| Symptom | Likely Cause | Investigation | Resolution |
|---|---|---|---|
| Minutes per viewer drop | Weak opening / content pacing | View retention curve first 5 minutes | Coach creators; improve intro hook |
| FHD% collapse | Encoder misconfig / network congestion | Compare bitrate logs & region distribution | Failover CDN; adjust adaptive bitrate ladder |
| Peak concurrency errors | Under-provisioned capacity | Error rate vs concurrency timeline | Pre-scale instances; add autoscale rule |
| High flagged rate | Inadequate moderation coverage | Flag timeline vs moderator shifts | Increase staffing; refine prohibited terms |
| Low engagement rate | Passive consumption / UI friction | Chat open rate vs viewers | Promote chat prompts; highlight Q&A moments |
| Start failure rise | Client version regression | Compare failures by app version | Rollback / hotfix offending version |
| Rebuffer complaints spike | Regional network issue | Geo segmentation of quality metrics | Route traffic alternate edge; adjust CDN config |
Best Practices
Optimize First Minute
Optimize First Minute
Concentrate most compelling hook; retention early strongly correlates with total watch time.
Progressive Interactivity
Progressive Interactivity
Seed chat prompts at predictable cadence (e.g., every 3–5 minutes).
Creator Feedback Loop
Creator Feedback Loop
Share stream-specific leaderboard metrics with hosts for iterative improvement.
Intentional Scheduling
Intentional Scheduling
Stagger high-demand streams to flatten concurrency peaks.
Quality Monitoring
Quality Monitoring
Automate alert when FHD% drops & average bitrate declines concurrently.
Safety Automation
Safety Automation
Combine keyword + behavior heuristics to preempt escalation surges.
Quick Reference
Reach
Unique Viewers
Depth
Avg Minutes / Viewer
Engagement
(Chat + Reactions)/Viewer
Quality
FHD %
Capacity
Peak Concurrency
Safety
Flagged Rate
Related
Livestream Moderation
Manage live safety
Activity Analytics
User baseline
Social Insights
Qualitative sentiment
Raw Data Export
Deep dive data
Need advanced QoE metrics (e.g., rebuffer ratio, join latency) or monetization funnels? Contact support to enable extended telemetry.