$142
1 · Context · An eight-top earns $142 per table-hour, +159% versus a deuce
$/table-hour rises monotonically from the deuce up to the eight-top. When tables bind the floor, big parties dominate the day's yield.
$27.50
2 · Context · A deuce earns $27.50 per seat-hour, +86% versus an eight-top
$/seat-hour falls monotonically from the deuce down to the eight-top ($14.76). When seats or kitchen bind the floor, small parties dominate the day's yield.
+45% chairs
3 · Context · Chairs and kitchen bound the floor, not tables — +45% more guests through the same table count
Peak table concurrency ran ~20 active turns per store — essentially at the typical-Sunday P90 of 22. The lift wasn't more concurrent tables; it was +45% chairs filled at those tables, plus +18% more total turns. The kitchen worked 1.45× a typical Sunday against an FOH cut 11%.
+2 hours
4 · Didn't · The labor cut was mis-timed by two hours — peak ran 11 AM–1 PM, FOH was thinned 9–11 AM
FOH cut 11% and BOH cut 8% against a 9–11 AM window. Active turns peaked at 11 AM (322) and held within 2.5% through 1 PM (314) — three consecutive peak-equivalent hours, all outside the labor window. For 2027, extend FOH and BOH past the current 11 AM tail through 1 PM.
5Mother's Day moved the mix, not the curvesDeuces gave up 15 points of share (48.5% → 33.5%); five-plus parties nearly doubled (11% → 21%). But yield by bucket held within $1–$4 of typical Sunday rates — check size rose modestly, turn times ran 3–5 minutes longer, the two effects offset. The operator fed +18% more turns through the same floor by serving larger parties at the buckets it already runs well.−15 / +10 pts
6Demand fit, not discipline, drove the two top-yield winsBoth kept large-party share under 14% (vs 21% portfolio avg) and turned floors 30–38% harder than typical — the portfolio's top YoY (+7.7% each). The mechanism wasn't a unique host-stand model: at Easter 2026 the top-yield store combined tables on the fly like every other store. Their Mother's Day cleanness came from a demand mix that fit standalone capacity.12% vs 35%
7The day got family-heavier in the back half of serviceAverage party rose from 2.90 at 8 AM to 3.71 at 11 AM, dipped to 3.51 at noon, then peaked at 3.92 at 1 PM. Family demand concentrates in the post-church 12–1 PM window, not the 9–10 AM early-bird wave. Family-magnet stores should pre-stage combinations for the back half of service with a planned second turn for the 11:45 family arrival.+1.02 guests
8Four stores ran the day as family destinationsEach ran 25%+ large-party share and 40%+ of dine-in revenue from five-plus parties. Floor-plan evidence shows the fix at three is pre-staging combinations the night before, not new inventory. The fourth is the structural exception — a 28-table dining room too small for the demand it receives.4 stores
9The structural-exception store has the worst demand-vs-capacity mismatch in the portfolioSmallest dining room (28 tables) absorbing 35.1% large-party share — the largest mix shift in the portfolio (+1.40 guests). Posted +3.6% YoY but Ovation 3.36 / 50% issue rate — the worst guest signal on the day. Table 24 hosted a single 10-top for 94 minutes and produced exactly one turn. Reservation pacing helps; the deeper question is whether the family demand is structurally too thick for the building.+1.40 guests
10The overflow store shows the demand-vs-capacity failure mode, mediated by improvised combinationHighest 5+-party revenue share (52%) but turn lift of only +14% (vs the top-yield store's +38%) and −4.8% YoY on a nearly identical floor. Five standalone big-tops can absorb ~25 large parties over the peak; the store received 58. The overflow landed in combination zones improvised in service (62+61+63, 31+32, 43+42). The fix is pre-staging those combinations the night before — an 8-position grid (5 standalone + 3 pre-combined) for reservation pacing.−4.8% YoY