Two-store analysis of the 5am–11pm → 5am–1am schedule change. Focuses on the 2 added hours (11pm and midnight) and their net contribution after labor.
Analysis date: March 18, 2026
Data sources: TimescaleDB — store_pl_monthly, daily_store_metrics, transactions, transaction_lines
Hours change: June 21, 2025 — added 11pm and midnight hours Status: Extended hours believed to still be in place
The extended hours at Francis are at best a break-even proposition. The two new hours generate approximately $111/day in truly incremental revenue ($77 uplift at 11pm + $34 uplift at midnight), yielding roughly $1,080/month in margin at the store's 32% merchandise rate. Against an estimated $3,200–$4,100/month in incremental labor, the coverage ratio is 0.30 on the conservative (incremental-only) basis. Under the most generous interpretation — crediting all late-night revenue to the extended hours decision — the ratio reaches 1.03, roughly break-even. The trend is declining from the summer peak into fall.
| Hour | Period | Txn/Day | Revenue/Day |
|---|---|---|---|
| 21 (9pm) — shoulder | Before | 33.8 | $748 |
| 21 (9pm) — shoulder | After | 33.6 | $737 |
| 22 (10pm) — shoulder | Before | 25.4 | $511 |
| 22 (10pm) — shoulder | After | 25.8 | $528 |
| 23 (11pm) — NEW | Before | 12.1 | $233 |
| 23 (11pm) — NEW | After | 16.4 | $310 |
| 0 (midnight) — NEW | Before | 1.4 | $43 |
| 0 (midnight) — NEW | After | 4.7 | $77 |
Key finding: Hour 23 already had significant pre-change activity — 12.1 transactions/day generating $233/day. These were likely last-call transactions before the old 11pm close. After the change, hour 23 increased to 16.4 txn/day (+$77/day). Midnight went from near-zero to $77/day (+$34/day).
Truly incremental daily revenue from the 2 new hours: ~$111/day.
| Month | Hr 23 Rev/Day | Hr 0 Rev/Day | Combined/Day |
|---|---|---|---|
| Jul | $352 | $76 | $428 |
| Aug | $356 | $82 | $438 |
| Sep | $294 | $77 | $371 |
| Oct | $292 | $79 | $371 |
| Nov | $225 | $75 | $300 |
Revenue peaked in July–August (~$430/day) and has declined to ~$300–$370/day in fall — likely seasonal.
Revenue in the new hours is 100% merchandise — zero foodservice. Prepared food operations are shut down by 11pm.
| Month | Employee Cost | vs June Baseline |
|---|---|---|
| 2025 May (last pre-change) | $33,288 | — |
| 2025 Jun (split month, change 6/21) | $30,023 | baseline |
| 2025 Jul | $38,305 | +$8,282 |
| 2025 Aug | $30,644 | +$621 |
| 2025 Sep | $29,824 | -$199 |
| 2025 Oct | $31,389 | +$1,366 |
| 2025 Nov | $33,468 | +$3,445 |
| 2025 Dec | $34,585 | +$4,562 |
Average monthly employee cost:
- Pre-change (Jan–May 2025): $28,933
- Post-change full months (Jul–Dec 2025): $33,036
- Difference: +$4,103/month
Cross-check: 2 added hours / 18 operating hours = 11.1% × $28,933 = ~$3,211/month theoretical increment. The actual $4,103 is somewhat higher, possibly reflecting overtime premiums.
| Component | Daily | Monthly |
|---|---|---|
| Hr 23 incremental revenue | $77 | $2,341 |
| Hr 0 incremental revenue | $34 | $1,034 |
| Total incremental revenue | $111 | $3,375 |
| Margin at 32% | $36 | $1,080 |
| Estimated incremental labor | ($3,200–$4,100) | |
| Net profit/(loss) | ($2,120) to ($3,020) | |
| Coverage ratio | 0.30 |
| Component | Daily | Monthly |
|---|---|---|
| Hr 23 total revenue | $310 | $9,424 |
| Hr 0 total revenue | $77 | $2,341 |
| Total gross revenue | $387 | $11,765 |
| Margin at 32% | $124 | $3,765 |
| Estimated incremental labor | ($3,200–$4,100) | |
| Net profit/(loss) | ($335) to $565 | |
| Coverage ratio | 1.03 |
The truth lies between these two approaches. The conservative basis likely understates the value (some of the "pre-existing" hour-23 traffic would have shifted to hour 22 if the store closed at 11pm). The generous basis overstates it (the store would retain most of that hour-23 revenue even if it closed at 11pm — customers would simply come at 10:45pm instead of 11:15pm).
The extended hours are not demonstrably profitable. At best they break even.
- Pre-existing hour-23 activity — 12.1 txns/day existed before the change, complicating incremental attribution
- Seasonal confounding — Before period is winter/spring; after period starts in summer
- No foodservice in new hours — 100% merchandise at ~32% margin
- Employee cost volatility — Monthly costs range $24,583–$38,305 in 2025
- November partial data — Only 17–19 days of data
Hours change: July 9, 2025 — extended to 5am–1am Rollback: September 13–14, 2025 — returned to 5am–11pm Duration of experiment: 67 days
The extended hours experiment at Liberty Lake was unprofitable from its inception. The two added hours (11pm–1am) generated approximately $353/day in revenue (based on summer 2024 proxy data), yielding roughly $124/day in gross profit at the store's ~35% merchandise margin. However, the incremental labor cost was approximately $152/day in July — meaning the store lost an estimated $28/day from Day 1. Over the 67-day experiment, the total estimated loss was approximately $1,600. The first actionable P&L signal appeared mid-August when July books closed showing employee costs at $31,561 — the highest month of 2025.
| Date | Event |
|---|---|
| Jul 9, 2025 | Extended hours begin (5am–1am) |
| Jul 31, 2025 | July P&L closes: employee costs $31,561, +18% vs. June |
| ~Aug 15, 2025 | July P&L likely available to management — first formal financial signal |
| Sep 13–14, 2025 | Hours rolled back to 5am–11pm |
| Month | Employee Costs | vs. June | EBITDA |
|---|---|---|---|
| May 2025 | $23,209 | -$3,637 | $36,956 |
| Jun 2025 (baseline) | $26,846 | — | $26,119 |
| Jul 2025 (extended from 7/9) | $31,561 | +$4,715 (+18%) | $27,699 |
| Aug 2025 (full extended month) | $26,872 | +$26 (flat) | $31,481 |
| Sep 2025 (rollback ~9/13) | $23,848 | -$2,998 | $20,632 |
| Oct 2025 (post-rollback) | $25,220 | -$1,626 | $39,750 |
Comparable PNW stores also spiked in July:
| Store | June → July | Change |
|---|---|---|
| 62111 Liberty Lake | $26,846 → $31,561 | +$4,715 (+18%) |
| 62102 Maple | $30,202 → $33,812 | +$3,610 (+12%) |
| 62113 South Hill | $24,974 → $27,787 | +$2,813 (+11%) |
Liberty Lake's premium over the comp average (+12%) suggests approximately $1,500 of the July labor increase was attributable to extended hours, with ~$3,200 being seasonal.
August costs fell back to $26,872 despite being a full month of extended hours — likely the store adjusted staffing (fewer overnight staff, restructured shifts) to reduce costs.
Note: Transaction-level data is only available through January 30, 2025. This analysis uses summer 2024 (Jul–Sep 2024) hourly data as the proxy for what the 2025 extended hours generated.
| Hour | Txns/Day | Revenue/Day | Avg Basket |
|---|---|---|---|
| 8pm (20) | 29.5 | $571 | $19.34 |
| 9pm (21) | 24.8 | $502 | $20.25 |
| 10pm (22) | 19.9 | $367 | $18.38 |
| 11pm (23) | 15.1 | $245 | $16.24 |
| 12am (24) | 6.3 | $108 | $17.15 |
Combined: 21.4 txns/day, $353/day — a steep dropoff from the 10pm hour.
| Day | Combined Rev/Day | Profitable? |
|---|---|---|
| Monday | $232 | No |
| Tuesday | $281 | No |
| Wednesday | $333 | Marginal |
| Thursday | $323 | Marginal |
| Friday | $433 | Marginal |
| Saturday | $459 | Marginal |
| Sunday | $431 | Marginal |
Monday and Tuesday were clearly unprofitable. Even Friday–Sunday were marginal after full labor costs.
| Component | Value | Basis |
|---|---|---|
| June 2025 employee costs | $26,846/mo | P&L actual |
| Hourly labor rate (18 operating hrs) | $49.70/hr | $26,846 / 30 / 18 |
| 2 extra hours at baseline rate | $99/day | |
| July actual premium over comps | ~$53/day | $1,500 / 22 extended days + $99 |
| Estimated daily labor for 2 hrs | $152/day | |
| Revenue in hours 23+24 | $353/day | Summer 2024 proxy |
| Gross profit at 35% margin | $124/day | |
| Daily net | -$28/day | Unprofitable from Day 1 |
Breakeven required $434/day in late-night revenue. Historical data showed only $353/day.
A pre-launch analysis using the store's own 2024 summer data would have shown:
- Late-night revenue: $353/day
- At 35% margin: $124/day gross profit
- Minimum staffing (1 clerk + overlap) at WA minimum wage ($16.66/hr + 30% burden = $43/hr) for 2 hours: $86/day
- Even at bare minimum single-person staffing, margin was only $38/day
- Any need for 2 staff members made it immediately unprofitable
This signal was available before the experiment started.
If daily transaction monitoring had been in place:
- Monday/Tuesday with <$300 combined revenue (below any breakeven)
- Average basket $16–17 (impulse purchases)
- Only 15–21 transactions combined for both hours
Available 1 week into the experiment.
July employee costs of $31,561 — highest month of 2025, +$4,715 over June. Combined with low late-night foot traffic observations.
First formal financial signal, available ~5 weeks after launch.
| Scenario | Days Extended | Est. Total Loss | Savings vs. Actual |
|---|---|---|---|
| Actual (rolled back Sep 13) | 67 days | ~$1,600 | — |
| Rolled back after 2 weeks (Jul 23) | 14 days | ~$392 | ~$1,200 |
| Rolled back after July P&L (Aug 15) | 37 days | ~$925 | ~$675 |
| Pre-launch analysis (never started) | 0 days | $0 | ~$1,600 |
The absolute dollar savings from earlier rollback ($675–$1,200) are modest. The more significant cost was opportunity cost — management attention, employee schedule disruption, and distraction from higher-impact improvements at a store generating $490K–$514K/month.
| Month | Total Revenue | Merch GP | EBITDA |
|---|---|---|---|
| May 2025 | $476,510 | $42,098 | $36,956 |
| Jun 2025 | $479,041 | $43,470 | $26,119 |
| Jul 2025 | $513,627 | $45,752 | $27,699 |
| Aug 2025 | $490,738 | $45,213 | $31,481 |
| Sep 2025 | $452,603 | $39,311 | $20,632 |
| Oct 2025 | $451,019 | $42,682 | $39,750 |
July had the highest revenue of the year but only middling EBITDA because employee costs ate the gains. October — the first full post-rollback month — had the highest EBITDA of the year ($39,750) despite lower revenue.
- Transaction data gap — Hourly data only available through Jan 30, 2025. Summer 2024 used as proxy for the actual Jul–Sep 2025 extended hours period. Actual late-night traffic may have differed.
- Proxy validity — Summer 2024 is reasonable: same seasonal period, similar merchandise revenue ($129K–$138K/mo in 2024 vs. $127K–$132K in 2025), same operating model.
- P&L monthly granularity — Cannot distinguish which hours generated which costs within a month.
- Staffing model unknown — Actual overnight staffing configuration (1 vs. 2 people, shift structure, overtime) is not captured in P&L data.
- Late-night margin assumption — 35% merchandise margin applied uniformly. Late-night purchases (beer, snacks, tobacco) may have different margins.
| Metric | Francis (62137) | Liberty Lake (62111) |
|---|---|---|
| Change Date | June 21, 2025 | July 9, 2025 |
| Hours Added | 11pm + midnight | 11pm + midnight |
| Status | Still extended | Rolled back Sep 13–14 |
| New-Hours Revenue/Day | ~$387 (gross) | ~$353 (proxy) |
| Incremental Revenue/Day | ~$111 | ~$353 (no pre-existing baseline) |
| Margin Rate | 32% | 35% |
| Monthly Margin from New Hrs | $1,080–$3,765 | ~$3,770 |
| Monthly Labor Increment | $3,200–$4,100 | ~$4,600 (July) |
| Coverage Ratio | 0.30–1.03 | ~0.82 |
| Verdict | Break-even at best | Unprofitable from Day 1 |
| Foodservice in New Hours | None | None |
| Declining Trend? | Yes (summer→fall) | N/A (rolled back) |
- Neither store demonstrates clear profitability from the 2-hour extension
- Zero foodservice revenue in the added hours limits margin potential to merchandise-only (~32–35%)
- Monday/Tuesday are the weakest days — if extended hours were to continue, consider weekend-only extension
- The pre-launch signal was available at both stores — historical hourly transaction data showed marginal economics before the experiment began
- October EBITDA at Liberty Lake ($39,750) — the highest of the year — came after the rollback, suggesting the extended hours were a net drag on profitability