feat(skills): add marketing skills collection

shared/skills/events/references/sponsorship-roi.md

@@ -0,0 +1,57 @@
+# Sponsorship & Exhibiting — Evaluate, Negotiate, Work the Floor, Follow Up
+
+Sponsorships are the most expensive way to do event marketing and the easiest to waste. The discipline: treat every sponsorship as a paid-acquisition channel with a cost-per-qualified-meeting, and make it beat your alternatives or don't buy it.
+
+## Should We Sponsor? (the evaluation)
+
+Run this before looking at the prospectus pricing:
+
+1. **Audience–ICP overlap, in absolute numbers.** Not "5,000 attendees" but *how many attendees are your buyer*. Ask organizers for the attendee breakdown by role/company type; check last year's attendee/speaker lists and social chatter. A 5,000-person event with 200 ICP attendees is a 200-person event for you.
+2. **Do the meetings math backwards.** Realistic qualified conversations = a small fraction of ICP attendees (a well-worked booth might convert 10–20% of relevant walk-bys into real conversations — directional, varies wildly by event). Then: total cost (sponsorship + travel + staff time + booth build) ÷ expected qualified meetings = **cost per qualified meeting**. Compare against what a meeting costs you from outbound or ads. If the event is 3× your outbound cost with no strategic upside, pass.
+3. **Strategic multipliers that justify a premium**: your exact buyers concentrated nowhere else, a category-defining event where absence is conspicuous (late-stage), or access you genuinely can't buy elsewhere (exec attendees who ignore cold outreach).
+4. **The counterfactual check**: what would the same budget produce in your best-performing channel? Sponsorship must beat that, not zero.
+
+Red flags in prospectuses: attendee counts without composition, "impressions" as the headline metric, leads defined as badge scans, and last year's sponsor logos heavy on companies that didn't return.
+
+## Negotiation: The Prospectus Is a Starting Point
+
+Sponsorship pricing is soft, especially inside 8 weeks. What to negotiate for (in rough order of value):
+
+1. **A speaking or panel slot** — worth more than a bigger booth; stage time converts better than floor space (see [speaking.md](speaking.md)).
+2. **Side-event rights** — permission/space to host a dinner, breakfast, or workshop for a curated list during the event.
+3. **Attendee list reality check** — full lists are increasingly rare (privacy); negotiate for opt-in scans, the registration-page question, or sponsored-session registrant lists. Get what's actually deliverable in writing.
+4. **Placement and timing** — booth position near traffic (coffee, entrances, main stage exit); demo-day timing if the event has one.
+5. **Price** — last, after the package is right. Unsold inventory close to the date discounts heavily.
+
+## The Side-Event Play (often better than the booth)
+
+The highest-leverage move in field marketing: skip or downgrade the booth, and **host a curated dinner or breakfast adjacent to the conference**.
+
+- 8–14 seats, hand-picked ICP attendees + a couple of magnetic guests (a respected practitioner draws acceptances)
+- Invite via personal outreach 2–4 weeks out (→ **cold-email** for craft); "join 10 [role]s for dinner during [event]" converts far better than any booth pull
+- No pitch. The host halo and the conversations are the product; follow-up carries the commercial weight
+- Economics: a dinner typically costs a fraction of a mid-tier sponsorship and produces *deeper* meetings with *chosen* accounts. This is also the play when you can't afford (or aren't allowed) to sponsor at all — you don't need the event's permission for your own dinner across the street.
+
+## Working the Booth (if you buy one)
+
+- **Staff it with people who can qualify and demo**, not whoever was free. Two energetic people beat five tired ones; write a shift schedule — floor fatigue is real and visible.
+- **A 30-second qualifying question** beats a pitch: "what does your team use for X today?" sorts buyers from swag collectors instantly. Have a graceful fast exit for non-ICP traffic.
+- **Capture context, not just scans**: after every real conversation, 15 seconds of notes — what they said, what they care about, the agreed next step. Voice memo or CRM app, same-hour. This is the raw material of follow-up that converts; a bare badge scan is a name with amnesia.
+- Demo stations for depth, one clear message on the booth itself (the category problem, not your feature list), and book-a-meeting QR that goes to a calendar, not a form.
+- **Book meetings before the event** with target attendees — the booth is a venue for pre-booked meetings, not just a net for walk-bys.
+
+## Follow-Up: Where the Sponsorship Is Won or Lost
+
+- **24–48 hour SLA**, tiered:
+  - **Hot** (real conversation, next step agreed): personal email referencing the conversation, calendar link, same or next day.
+  - **Warm** (conversation, no commitment): personal note + one relevant asset matched to what they said.
+  - **Scan-only**: one light "we were both at [event]" touch or nothing. Never dump scans into a sales sequence — it burns domain reputation and brand on people who don't remember you (→ **revops** for routing/scoring).
+- Whoever worked the booth writes or reviews the follow-up — the context lives in their heads and their notes.
+- Sequence the not-nows into nurture with event source tags (→ **emails**).
+
+## Measuring the Sponsorship
+
+- Log every touched contact with an event source tag; measure **qualified conversations → meetings → opportunities → pipeline → closed-won influenced**, on an influence window that matches your sales cycle (90 days is common for B2B; long cycles need longer windows).
+- Report **cost per qualified meeting and cost per opportunity** against your other channels — this is the renewal decision for next year, made with data instead of vibes.
+- Self-reported attribution ("met you at [event]") catches influence that source tags miss (→ **attribution**); badge-scan counts and booth traffic are activity, not outcomes — track for logistics, never report as results.
+- Judge a first-time event against a discount: your team's first run of any event underperforms its potential. A promising-but-unprofitable first year is a redesign signal, not necessarily a no.