The Outreach Runs page now tells the full story of every discovery: per-creator drop reasons in the Discovered Creators table, a funnel banner above the table, sortable + filterable columns, a rolled-up Summary row across all four stages, and a separate label for creators we couldn't verify under strict mode. No more guessing what happened between Discovery starting and your drafts appearing. Six visible improvements landed across the Outreach Runs page so you can audit and tune your discovery pipeline directly from the UI: Per-creator drop reasons — Every creator that discovery returned but didn't promote now shows exactly why — Below follower threshold, Below engagement threshold, Couldn't be verified, Missing content evidence, or Unresolved profile. Open the Discovered Creators table on any run; the chip on each row tells the story. Hover for the rule that applied. Funnel banner above the table — A new banner above the Discovered Creators table summarizes the run — '5 of 14 creators passed eligibility for scoring' — with per-platform breakdowns and direct links into Setup to adjust the threshold that filtered each bucket. You can see at a glance whether to relax follower minimums, tune engagement, or look at strict-mode coverage. Sortable + filterable Discovered Creators table — Click any column header (Creator, Engagement, Followers, Keyword, Stage, Evidence) to sort. Two quick-filter chips above the table let you toggle 'Engagement ≥ minimum' and 'Followers ≥ minimum' to focus on creators who passed those gates. The 'Filtered out' badge expands into a multi-criterion tooltip showing exactly which thresholds the row passed (✓) and failed (✗). Summary row across all four stages — Every Outreach Runs execution card now ends with a Summary row that rolls up Discovery → Eligibility → Scoring → Drafting into a single line — counts at each transition, total wall-clock duration, and the result. Click to expand for a per-stage breakdown including the eligibility drop histogram (followers / engagement / unverifiable / missing evidence / unresolved). Strict-mode drops named correctly — Creators we drop because we couldn't reliably measure their engagement (single evidence sample under strict mode) now surface under their own bucket — 'Couldn't be verified (insufficient evidence)' — instead of being lumped in with 'Below engagement threshold'. The tooltip points you at the strict-mode toggle in Setup so you can choose to admit those creators if you'd rather see them. Stage label clarity + Summary row polish — The post-scoring chip 'Scored' was misleading (it actually meant 'filtered out by score threshold'); it now reads 'Filtered out'. The Summary row's Started/Ended cells now match the per-stage row times exactly (UTC parity), the status badge stays on 'In progress' until all four stages complete (instead of flipping to Completed when only Discovery finishes), and an in-flight stage now shows '(running)' or '(pending Apify)' instead of an empty Ended cell.