skills/events/references/webinar-funnel.md
The webinar is the flagship hosted format because it's the whole event arc in miniature, repeatable monthly, and every stage is measurable. Work the four stages in order — each stage's conversion rate is a separate lever with separate fixes.
The funnel at a glance (typical B2B ranges — directional benchmarks, not targets; your own trend line is the real baseline):
| Stage | Metric | Typical range |
|---|---|---|
| Registration page | Visitor → registrant | 30–50% (warm traffic), 10–25% (cold) |
| Show-up | Registrant → attendee | 35–45% live; lower for cold/ads traffic |
| Hold | Attendee stays past minute 40 | 50–70% |
| Convert | Attendee → next step (trial, demo, offer) | 5–15% of attendees for a soft CTA; 1–5% direct purchase |
The topic does the targeting and most of the selling:
The registration page is a landing page (→ copywriting for craft); webinar-specific rules:
The promo plan — start 2 weeks out, not 6 (urgency compresses better than it stretches):
Registrants are cheap; attendance is the funnel's leakiest joint. The show-up system:
The arc that converts without feeling like a pitch:
Hold-rate mechanics throughout: deliver on a specific promise made in minute 1 at minute ~35 (announced), use pattern breaks every ~7 minutes (poll, story, screen change), and never front-load housekeeping.
Segment by behavior, then sequence (→ emails for craft):
| Segment | Play |
|---|---|
| Attended, engaged (stayed for offer, asked questions) | Personal follow-up within 24h referencing their question; direct next step |
| Attended, left early | Replay + timestamp to what they missed; softer CTA |
| No-show | "Sorry we missed you" + replay with a deadline. No-shows are warm — they raised their hand once; a 2–3 email replay sequence recovers a meaningful fraction |
| Replay-registrants | Same as no-shows, minus the apology |
Skill category identified via 2026-07 competitive research (webinar-marketing in alirezarezvani/claude-skills, MIT — idea credited; content authored from scratch to this repo's standard). Benchmarks are directional industry ranges — treat your own trend line as the baseline.