TL;DR: Four tabs at the top — Overview, ICP, Signal Source, Engagement Stage. Each tab is a table of cohorts × outcome metrics. Color-coding flags above-average and below-average cells.
Overview tab
The default landing view. Big numbers at the top:
- Total qualified records (current window)
- Total enriched
- Total synced to destinations
- Total replied
- Total meetings booked
- Total signed up
- Total purchases
- Total LTV
Each carries a delta vs the previous equal-length window (last 30 days vs the 30 before that, by default). Green delta = improving; red = declining.
Below the totals, a stacked area chart shows the funnel over time — captured → qualified → enriched → synced → replied → meeting → signup → purchase. The width of each band reflects volume; the slope reflects drop-off rate.
ICP tab
One row per ICP. Columns: ICP code, qualified count, enriched %, synced %, replied %, meeting %, signup %, purchase %, LTV, avg credit cost per purchase.
Color coding: cells above the workspace average get a green tint; cells below get a red tint. The intensity scales with the deviation from average.
What to look for:
- ICPs at the top of purchase % are your money lanes. Investigate what makes them work — usually a tighter Persona definition or a higher-converting signal source.
- ICPs with strong reply rate but weak purchase rate — somewhere in the funnel between meeting and purchase, you're losing them. Outside Unstuck Engine's control directly, but the signal-and-message combo got them to the table.
- ICPs with high qualified count and low replied % — signal noise. The ICP is qualifying too broadly. Tighten Filter parameters.
Signal Source tab
One row per signal type. Same column structure. Often the most actionable view.
- Signals at the top of reply rate — your highest-converting source. Increase the signal's autoqualify confidence; consider relaxing other signal sources to feed the audiences that consume this one.
- Signals at the bottom — either noise or context-dependent. Investigate before pausing; sometimes a low-reply-rate signal is still the best leading indicator for a longer sales cycle.
Click a row → drill-down to per-rule performance (relevant for signals like Keywords with multiple rules per signal). You see which specific Keywords rule is producing replies vs noise.
Engagement Stage tab
Four rows — Cool, Warm, In-market, Hot. Performance metrics at each Stage.
The pattern across most workspaces:
- Hot — highest reply rate, smallest volume
- In-market — second-highest, larger volume
- Warm — workhorse stage, biggest volume
- Cool — low reply but sometimes surprisingly high purchase rate (long sales cycles)
The Cool row often surprises. Workspaces that filter out Cool records miss the longer-cycle conversions.
Filtering and time range
The toolbar at the top of every Insights view:
- Time range — Last 30 / 90 days / All time / Custom range
- Filter — restrict to specific ICPs, Personas, signal sources, audiences
- Group by — overlay an additional grouping (e.g. group ICP tab by Persona)
- Compare to — overlay a comparison period (default is the previous equal-length window)
Exporting
Top-right of the dashboard: Export. Generates a CSV with the current view's data. Useful for QBR decks and external reporting.
What's NOT in Insights
- Per-rep performance. Insights is workspace-level, not per-user. Per-rep reporting lives in your sequencer / CRM.
- Individual record outcomes. Click into a record (from Leads / Accounts) to see its specific outcome history. Insights aggregates; the detail pages narrate.
- Real-time alerts. Insights is reporting. For "X happened just now" notifications, use the notifications system (see Settings → Notifications).