Forecast gap, Tuesday
40% empty
19 of 48 covers unbooked at Monday noon
On Monday the model saw Tuesday dinner trending 40% empty. By the time the room opened, it had already quietly invited the right regulars — and held back the ones who'd have walked in anyway. Read by Eleanor Marlowe, Owner
A forecast that only tells you the room will be quiet is a worry, not a tool. This one is wired to do something about it — but only within the gap it can prove, and only with guests who've consented to hear from us.
It stacks three things the house already runs: the demand forecast, the guest book, and the messaging channels. None of it is new plumbing — it's the existing parts, pointed at one job.
Tuesday flagged 40% empty against capacity.
Rank regulars by spend, recency & preference.
One Tuesday-only invitation, in the house voice.
SMS & email to the consented few — not the list.
Opens, bookings, seats filled by Tuesday noon.
Drawn from the guest book: regulars with three or more visits, a spend at or above the $80 average, marketing consent on file, and a preference that suits a quiet Tuesday. Two examples:
"A quiet Tuesday's our favourite kind. Chef's table this Tuesday only — two glasses poured to suit your table, on us. Marcus, the Manuka-smoked duck and that Bridge Road porter will be waiting."
Nine bookings, average party 1.8, filled 16 of the 19 empty seats — the Tuesday gap closed from 40% to 6% without touching the door price for anyone who'd already planned to come.
The first draft segment was bigger — 63 regulars. It swept up loyal, full-price guests who book Tuesdays unprompted and have never needed an incentive. Discounting them fills no empty seat; it just hands away margin on covers we already had. The guard pulled them out of the send and explained why.
Wei Lim — books a quiet Tuesday roughly monthly, always full price, opened the last three newsletters without an offer. Would have come anyway. Inviting at a discount costs ~$22 in margin and fills no incremental seat.
Grace Goh — top-decile spender, no price sensitivity in her history, already a confirmed Tuesday booking on the sheet. Already coming. The offer would have discounted a seat that was never empty.
Marcus Tan — a true regular, but hasn't been in for seven weeks and only ever on busy nights. Genuinely incremental. A tailored Tuesday nudge wins back a cover the room would otherwise lose.
The rule: the offer goes only to guests whose presence is uncertain on the night with the gap — never to a confirmed booking, and never to a reliable full-price regular. Of 63 candidates, 13 were held. That protected an estimated $310 of margin that a blanket "fill the room" send would have given away.
This page is a concept. The numbers are hand-built to tell one Tuesday's story honestly; nothing here ran against a live room. What it illustrates is real, though, and it's the point of the expansion phase: the three use-cases already on the table — the demand forecast, the guest book, and the messaging channels — stack into a single action. The forecast spots the gap, the guest book picks who to ask, the channels do the asking, and a margin guard sits across the whole thing so it can't give away the room to fill it. Today that orchestration is a sketch; the parts beneath it are further along.