Strategy and event selection

Start from the objective, not the event

Events get proposed backwards — a prospectus lands, a peer is going, a founder likes the city. Reverse it. The annual event portfolio is derived from BU goals:

One event can serve two objectives; it cannot serve four. If the objective column is blank, the answer is no.

The evaluation method

The role SKILL.md's five questions (audience match, cost to opportunity, comparable alternatives, brand fit, kill condition) are the frame; the scorecard operationalises them. Rules that make it honest:

1. No evidence, no score. A criterion scored without evidence in the evidence block is scored 0. The sales deck for the event is marketing; the past attendee list is data.
2. ICP density is the heaviest weight (25/100) on purpose. A 300-person room at 30% ICP beats a 5,000-person hall at 2%. Compute it: named-org overlap with our target-account list, titles from the prior-year list, sponsor roster as a proxy for who buys access to that audience.
3. Access matters as much as attendance (15/100). A perfect room you can't reach — no list, no meeting tool, side events banned, booth-traffic-only — is a billboard, not a pipeline instrument. Establish the access model before scoring: attendee list timing, official meeting platform, whether the venue allows hosted side events, hallway culture.
4. Score the mode, not just the event. The same conference can be a fail at sponsor tier and a pass as attend-and-hunt plus a hosted dinner nearby. The scorecard's mode recommendation is the actual output; "go cheap" is a first-class answer.
5. Pass loudly. A passed scorecard is filed in data/events/ with its score so next year's decision starts from evidence, not vibes.

Where events come from (scouting inputs)

• The Researcher's Monday Week Ahead (industry events section) and news-desk signals.
• Client + prospect calendars: ask in QBRs and discovery calls what they attend — the highest-signal input there is.
• Inbound sponsorship pitches — always through sponsorship-evaluation.md first; never evaluated on the sales call (role SKILL.md).
• Speaker-CFP calendars for target stages (CFPs close 3–6 months out — scout two quarters ahead).
• Peer observation: where do credible non-competing firms selling to our ICP keep showing up twice? Repeat presence implies it paid back once.

Maintain the rolling register: every scouted event gets a row in ${INTO_OPS_REPO}/data/events/register.md (event, date, CFP deadline, early-bird deadline, scorecard link, state). The register is scanned in the CMO's weekly routine so deadlines drive decisions, not the other way round.

Target audience definition (per event)

The brief names the audience precisely, from the BU file and brand file — for Consulting: founders/CEOs/COOs/CTOs and Heads of Product/Ops/CS at mid-market B2B, plus software companies adding AI to their product. Then localise it to the event: which named accounts and which named people are plausibly in this room? The comms plan (06) and sales pack (07) both key off that named list — building it is selection work, not prep work. If you can't name 20 target people/orgs for a conference (or 12 for a dinner), the event fails the access test.

Timing + portfolio discipline

• Decide big commitments at T-16w or earlier (early-bird pricing, good booth positions, speaker slots all expire earlier than teams expect — chapter 05).
• Cap concurrency: one live event per month per BU is the ceiling while the team is small — follow-up capacity is the binding constraint, not travel budget (doctrine rule 5).
• Sequence for compounding: a dinner series in one city each quarter beats four one-off cities. Relationships compound where you return.

Kill conditions and withdrawal

Every scorecard names its pre-commit kill criteria; the budget sheet (03) records the last-refund date. After commitment, withdraw when a kill criterion trips and the refund window is open — sunk cost is not a reason to attend. If the window is closed, downgrade the mode: cut the booth build, halve the travelling team, keep the pre-booked meetings. Log the withdrawal decision like any other (TEAM.md §5).