event-account-plan

Applies once an attendee, registration, or target-audience list exists and there is still time to act on it — typically the two weeks before the show. Produces a ranked account file, a decision-maker file, and a war-room page: which companies to work, the buying committee behind each, and the play per segment. Not a summary of the event.

Collect five things

Ask only for what you cannot find yourself. Missing 1 or 2 blocks the work.

1. The attendee list — export, sheet, or paste. Company rows, and named registrants if you are lucky.
2. Whose GTM this is — the vendor's site. Homepage, product pages, and the customers/logos page.
3. The event — name, dates, venue, attendee count, track list. The track list tells you which personas actually show up.
4. Credit policy — ask explicitly whether you may spend enrichment credits. Most users say no. Default to search-only and say so in the deliverable.
5. How to treat existing customers — expansion track (the usual answer) or excluded. Never silently drop them.

The pipeline

Each step feeds the next and is cheap to redo alone.

1. Normalize. Parse to two tables — accounts and named attendees. Dedupe accounts by parent company, not by row; big lists are full of divisions registering separately. Merge them, keep every division name in a sheet_names field, and count them: multiple divisions registering independently is a buying signal, so carry the count into scoring. Check which columns are actually populated first — a location column that is empty on every row is common.
2. Read the ICP before scoring anything. From the vendor's own site: what the product does, the named buyer personas, the verticals, the deal-blocking compliance regimes, the competitors they name, and the customer logo list. Skip this and your weights are guesses.
3. Resolve companies. Run accounts through whatever company-data connector is available for ID, revenue, employee count, HQ, and industry. Batch by domain. Expect roughly 90% to match; carry the misses through with blanks rather than dropping them.
4. Find signals. Firmographics rank accounts; signals tell you why now. Pull tech-stack tags for the vendor's own product and for every named competitor. A dead or dying competitor is the best signal that exists: if one has shut down, been acquired and sunset, or gone end-of-life, its install base must migrate. Search for that explicitly, every time. Read references/signal-sourcing.md before the first batch — it covers the connector traps that silently return zero rows.
5. Score and tier. Additive 0–100, weights adapted to the ICP from step 2, with a penalty for partner-motion segments. Full rubric, bands, and tier cuts in references/scoring-model.md. Score existing customers on a separate expansion track — a customer scores high on everything and will otherwise crowd out every new logo.
6. Find the buying committee. Tier 1 and top expansion accounts only. Names and titles are enough. Target the personas from step 2, cap at about eight per account, rank by persona fit first and seniority second.
7. Write the plays. One per track — migrate the defunct competitor, displace the live ones, new logo, expand customers, booth demo. Each gets a title, an audience, a body, a verbatim opening question a rep can say out loud, and a proof point. Then count buyer-persona titles among the named attendees: if most attendees do not own the budget, say so plainly and re-aim the copy at operational cost. See references/plays-and-messaging.md.
8. Ship all three deliverables — accounts file, decision-makers file, and the war-room page. Column lists, page sections, and the verification pass are in references/deliverables.md.