Analytics always scopes to exactly one connected Instagram account — there's no combined "all accounts" view. Switching accounts from the header keeps your current date range; it doesn't reset it. Returning later, it reopens on the account you last viewed.
Range and tabs
Choose 7, 30 or 90 days, or custom dates. The URL preserves the range.
- Summary — Conversations, qualification, meetings, closes, revenue, ticket, win rate, open pipeline, speed to meeting and a funnel.
- Leads — Pipeline distribution and attribution by source, channel, avatar, entry point and campaign where data exists.
- Improvements — AI-generated recommendations based on this period's real numbers; the badge counts the ones you haven't applied yet.
What the numbers actually mean
Win rate and open pipeline correctly leave out leads the AI disqualified before ever booking (a separate "No qualifies" state from LOST) — only a real booked-then-didn't-close outcome counts against win rate, and a disqualified lead no longer inflates the open pipeline count.
The funnel on the Summary tab is four simple counters — conversations → qualified → meetings → closed — not a step-by-step timeline of what happened to each lead.
Attribution (source, channel, avatar, entry point, campaign) is entirely automatic — it only shows up when Meta's own webhook included referral/ad metadata, or when the AI itself tagged a lead's avatar or channel during qualification. There's no manual tagging UI, so "when data exists" means exactly that: if nothing ever wrote it, that lead shows up as untagged, not missing.
Applying an Improvements recommendation isn't a suggestion you queue — it writes that instruction directly into your live AI agent's prompt immediately, org-wide, the same as editing AI Agent config yourself.
Bounded history
If a workspace passes 50,000 leads or 50,000 meetings, Analytics stops pulling older rows beyond that cap — the oldest ones drop first, not just ones outside your selected date range. A notice appears when this applies.