Analytics
Track calls, answer rate, billed minutes, spend, outcomes, agents, and hourly activity across any date range on the Analytics dashboard.
What you'll learn
What each analytics metric means, how period-over-period comparison works, how your timezone affects day and hour boundaries, and how to export the underlying data.
Analytics
The Analytics dashboard gives you a complete, at-a-glance view of your calling activity across any date range: how many calls ran, how many connected, how long they lasted, what they cost, why they ended, and when they happen. Use it to monitor performance, control spend, and spot issues.
Home defaults to the last 7 days; Analytics covers any range you select.
Choosing a date range
At the top of the dashboard, pick a range:
| Option | Description |
|---|---|
| 7 days | The last 7 days (default on Home) |
| 30 days | The last 30 days (default on Analytics) |
| 90 days | The last quarter |
| Custom | Any start and end date from the calendar |
Every metric — including the period-over-period delta chips — recalculates for the selected range. You can also filter the dashboard by direction (All / Inbound / Outbound).
Metrics
Every metric is calculated for the range you select, alongside its value for the previous period of equal length (for example, this week vs. last week) so you can see the direction of travel.
| Metric | What it measures |
|---|---|
| Total Calls | Every call created in the range |
| Answer Rate | Share of calls that connected (the recipient answered or the web client joined) |
| Connected | Count of calls that connected |
| Billed Minutes | Minutes billed for completed calls |
| Avg Duration | Average length of connected calls |
| Avg Latency | Average agent response startup time |
| Total Spend | Cost across the range (LLM + TTS + STT + telephony + platform) |
| Live Calls | Calls active right now (not range-bound) |
| Completed | Share of calls that ran cleanly enough to bill. Platform failures are excluded, so this sits at or below Answer Rate |
| Engaged | Share of connected calls lasting 30 seconds or more — early hangups drag it down |
| Human reached | Share of calls that did not end in voicemail, out of calls that reported an end reason |
| Active days | Share of days in the range on which at least one call ran |
Delta chips are color-aware: for latency, a decrease is shown as good; for volume and answer rate, an increase is good. When the previous period had no activity, the chip reads new instead of a misleading percentage.
Timezone
All dates and hours use your configured timezone (set in user settings). Day boundaries are calculated in that timezone, so "today's calls" reflects your local day, not UTC.
Understanding your metrics
| Indicator | What it means | Action |
|---|---|---|
| Low answer rate | Calls not connecting | Review number quality, calling times, and telephony setup |
| High voicemail / no-answer | Calling at the wrong times | Adjust campaign scheduling |
| Rising avg latency | Slower agent responses | Check model/provider choice and pipeline health |
| Spend climbing faster than calls | Cost per call increasing | Review model and voice choices in the cost breakdown |
| Calls concentrated outside business hours | Calls landing at low-engagement times | Retune campaign windows |
Data export
For bulk export across all call types, use the Call History API with date-range filters, or a campaign's runs endpoint for campaign-specific results.
Frequently asked questions
| Question | Answer |
|---|---|
| Can I break down by agent? | Yes — calls are broken down by agent for the range you select. |
| How far back does data go? | As far back as your call history is retained. Pick any custom range. |
| Are browser calls included? | Yes. Outbound, inbound, and WebRTC (browser) calls are all included; the direction filter separates them. |
| Do failed calls count against spend? | No. Billed minutes and spend count only completed calls; platform failures are never billed. |
| What timezone are dates in? | Your configured user timezone. All day and hour boundaries are calculated in it. |
Next steps
Call History
The Calls page — list view filters and columns, plus the per-call detail page with Transcript, Logs, Latency, Insights, Cost, and Config tabs. Every metric defined with what it means and how to use it.
Tracing
Integrate Langfuse for tracing LLM calls, tool invocations, and pipeline performance across every voice AI call.