Scheduling without excess-hours penalties
Build legal weekly coverage around demand peaks and employee restrictions.
Decision first
After any midweek edit, hover actual hours worked; the new grid does not erase shifts already completed.
A schedule is a seven-day assignment of compatible employees to required workstations inside opening hours, constrained by remembered weekly hours and weighted demands.
Set profitable opening hours first.
Complete operating playbook
- ARTICLE LENGTH
- 5,338 words
- DEEP-DIVE SECTIONS
- 17
- RESEARCH LEDGER
- 10
Scheduling model in Build 3674
A schedule is a seven-day assignment of compatible employees to required workstations inside opening hours, constrained by remembered weekly hours and weighted demands.
Treat opening hours, workstation coverage and employee status as separate layers. Opening a shop permits customers to arrive; it does not staff a register. Assigning a cashier to one register does not cover a second register, cleaning station or security equipment. Giving an employee forty hours does not prove those hours occupy a profitable or required station. The valid schedule is the intersection: each open hour has the intended number of staffed compatible stations, every employee remains within the status band, and individual day, time and equipment demands remain acceptable.
The current documented bands are 30–50 hours for full-time and 10–30 for part-time. Hours completed earlier in the week remain in the employee counter even after the future schedule is changed, and reset Monday morning. This is why a visually correct pattern can remain red until the next week. Build 3674 added clearer schedule tooltips, while EA 0.5 closed an exploit that erased worked hours through midweek editing. Always read Hours Worked before diagnosing a total.
EA 0.8 rebuilt auto-fill to consider demands, split shifts more sensibly, fill one day or the whole week, copy and paste days, rename workstations and search workers. Build 3014 then let the slower mode use shifts up to twelve hours and avoid tiny two-hour fragments, while Fast Auto-fill retained speed. Auto-fill is a feasible-roster assistant, not a profit optimizer: it cannot know that an empty late hour should be closed or that one extra noon cashier has high marginal value.
Action checklist
- Set profitable opening hours first.
- List required stations by hour.
- Respect 30–50 and 10–30 bands.
- Inspect remembered hours before editing.
- Audit auto-fill against demand and profit.
Build a coverage matrix
The fundamental calculation is required station-hours by role and time. Headcount is a result of this matrix, never the starting input.
Create rows for each hour and columns for each station type. Enter one where a station must operate and additional units where simultaneous demand justifies overlap. A store open 08:00–20:00 with one register requires twelve cashier-hours daily; adding a second register 12:00–15:00 adds three, not another twelve. Repeat for cleaning, security, specialist stations, factory machines and office desks. Sum by day and week, retaining time bands because two totals with different placement are not interchangeable.
Derive coverage from evidence. Use Insights, queues, lost-customer alerts, sales by hour where available, dirt accumulation, theft and production bottlenecks. Building customer capacity is not automatically a staffing target, and every installed desk need not be occupied. Add a station when its incremental gross contribution exceeds loaded labor and supporting cost. Remove or shorten a station after a comparable full-week test shows sustained underuse, not after one weather, rival or market-event day.
Mark hard constraints in the matrix: a No Weekends employee cannot cover Saturday or Sunday; a No Evening Shifts employee cannot fill a late block; a four-day demand limits distinct workdays even when total hours fit. Color uncovered cells only after applying compatibility. This prevents the false solution of having enough aggregate labor but no acceptable person for the exact Sunday evening gap. The hiring requisition is the unresolved cells grouped into a feasible employment band.
Action checklist
- Use hour-by-role cells.
- Count overlap as incremental station-hours.
- Base coverage on observed bottlenecks.
- Mark hard employee constraints.
- Recruit from unresolved cells.
Full-time and part-time arithmetic
Weekly employee bands convert the coverage matrix into a lower bound on workers, but demand constraints and fragmentation determine the real number.
For H required hours, the theoretical minimum full-time headcount is ceiling of H divided by fifty. That is only a lower bound: 98 hours suggests two workers, but two at forty-nine each may be impossible when the store is open seven days and either worker has a day-count or weekend restriction. A resilient design usually targets forty to forty-eight full-time hours, preserving room for edits and avoiding an exact maximum. Part-timers efficiently absorb residuals between ten and thirty.
Estimate headcount with target hours rather than maximum hours: full-time equivalents equal required role-hours divided by the chosen target, commonly forty to forty-five. Then solve the residual with compatible part-time staff. If 132 hours are required and the target is forty-four, three full-timers fit arithmetically. If forty-eight hours occur only on weekends, at least one worker must accept that placement; a weekday-only pool cannot solve it regardless of total capacity.
Avoid exactly thirty where possible because common help descriptions include the boundary in both ranges. The live employee tooltip decides, but a 29-hour part-timer and a 31-hour full-timer are more robust than two people balanced on thirty. Pay is hourly, yet benefit cost is per covered person and insurance has a ten-person minimum, so more part-timers can increase loaded cost. Compare wage and benefits per staffed hour, not wage alone.
Action checklist
- Use ceiling H/50 only as a lower bound.
- Plan full-timers near 40–48 hours.
- Use part-timers for 10–30-hour residuals.
- Avoid fragile thirty-hour boundaries.
- Include per-person benefits in roster cost.
Worked roster: 84-hour single station
One register open 08:00–20:00 every day requires eighty-four weekly hours. Two full-time workers can cover it with reasonable slack.
Roster A uses Worker A for 08:00–14:00 every day and Worker B for 14:00–20:00 every day. Each receives forty-two hours across seven days. It is simple and status-valid, but fails any four-day, five-day, No Weekend, No Morning or No Evening demand. Simplicity is useful only when the hired demand profiles support it. It also creates a single-point absence in each half-day without HR temporary coverage.
Roster B uses Worker A Monday–Friday 08:00–16:00 for forty hours, Worker B Monday–Friday 16:00–20:00 plus both weekend days 08:00–20:00 for forty-four hours. This fits totals but gives B long weekend shifts and seven distinct workdays. Build 3014 auto-fill supports shifts up to twelve hours, but an employee's own day and time demands still decide satisfaction. If B wants five days, split weekend coverage into a 24-hour part-time worker.
Roster C uses two full-timers at thirty-five hours and one part-timer at fourteen. A covers five seven-hour morning blocks, B covers five seven-hour late blocks, and C covers remaining compatible cells. It costs another benefit head but offers resilience and easier demand matching. Compare expected vacancy loss and insurance minimum effects. For an early store without HR, C may preserve opening continuity; at scale, A may be cheaper because HR covers sickness.
Action checklist
- Confirm 12 × 7 = 84 hours.
- Test distinct workday demands.
- Check twelve-hour shift acceptability.
- Price the third employee's benefit cost.
- Choose resilience deliberately.
Worked roster: 98-hour long-day store
A 08:00–22:00 seven-day station needs ninety-eight hours, exposing why two maximumed full-timers are an unsafe design.
Two workers at forty-nine hours can cover the raw total, but any incompatible day, illness, training interruption or midweek edit breaks the plan. A practical base roster assigns A 08:00–14:00 daily for forty-two, B 14:00–20:00 daily for forty-two, and part-timer C 20:00–22:00 daily for fourteen. This preserves status bands and makes the expensive edge visible as a separate business decision.
If evening contribution is weak, test closing at 20:00. The fourteen removed hours may eliminate the part-timer, their insurance allocation and management slot while losing little sales. If evening demand is strong, search specifically for a part-time worker without No Evening or No Weekend. Do not hire a generic full-timer and force thirty hours elsewhere merely to cover fourteen valuable edge hours; that turns a small schedule need into paid low-productivity coverage.
When a second register is needed 12:00–14:00 on weekdays, the added ten hours can raise C to twenty-four by moving part of their work to noon overlap. Verify no simultaneous conflict: one employee cannot staff the evening and noon at the same moment, but same-day separated blocks may be possible subject to F1 restrictions. Calculate incremental customers needed per overlap hour from loaded hourly cost divided by contribution per sale.
Action checklist
- Recognize two-at-49 as brittle.
- Separate edge coverage from base coverage.
- Test closing weak late hours.
- Search for exact residual compatibility.
- Calculate overlap break-even customers.
Worked roster: two-station peaks
A store open 09:00–21:00 daily needs one base station for eighty-four hours and a second station for measured forty-eight peak hours, totaling 132.
Three full-time employees at forty-four hours each match the total. A can cover 09:00–15:00 daily for forty-two plus two peak hours; B covers 15:00–21:00 daily for forty-two plus two peak hours; C covers the remaining forty-four peak hours. This arrangement is only illustrative because C's peak blocks may be fragmented across every day. Auto-fill's slower mode tries to avoid two-hour fragments, but a human may choose profitable fragments intentionally.
An alternative uses two forty-two-hour base workers and two twenty-four-hour part-time peak workers. It adds one person but separates operational roles cleanly and can satisfy No Weekend or day-count demands. Insurance and HR capacity rise, while paid idle time falls. Compare loaded weekly cost and the probability that a peak specialist's absence loses customers. If peak demand shifts by weekday, copy a base day and adjust only the affected cells rather than forcing a shared schedule.
Validate the second station through a matched test. Keep stock, price, marketing, opening hours and employee skill stable. Staff the second station during selected comparable peaks and record sales, customers served and lost-customer indicators. Incremental contribution equals additional units sold times unit margin minus incremental labor. A queue disappearing without higher sales can mean customers merely redistributed; do not assume every person touching the second register is incremental.
Action checklist
- Sum 84 base plus 48 peak hours.
- Compare three full-time with mixed staffing.
- Use per-day edits for shifting peaks.
- Control other variables during tests.
- Measure incremental contribution.
Demand-aware placement
A mathematically complete schedule can still drive resignations when days, times, equipment or employment status contradict employee demands.
Apply demands in priority order. First keep the Critical full-time or part-time band valid. Next satisfy restrictions that make a scheduled hour unacceptable: No Weekends, No Mornings, No Evenings and specific day counts. Then assign equipment-sensitive office workers to compliant desks. Insurance and shared amenities can be solved through HR groups and layout. Nice-to-have items may remain unmet if the weighted balance is positive, but document that decision and watch the trend.
Translate day-count demands literally. A worker scheduled for one hour on a sixth day has worked six days even if weekly hours are low. Translate time restrictions across every block; a mostly daytime schedule still violates No Evening when one closing shift crosses the prohibited period. Hover the live tooltip for exact definitions because labels alone do not expose the boundary. Use renamed workstations so desk demands are auditable rather than relying on visual memory in a large office.
Use a compatibility matrix with employees as rows and shift templates as columns. Mark valid, valid after cheap accommodation, and impossible. Assign the most constrained shift first: weekend evenings, specialized machines and unique HQ roles. Flexible employees fill the residual. This is the opposite of placing the highest-skill person first, a method that often strands an inflexible but adequate worker and forces another hire.
Action checklist
- Protect Critical employment status.
- Count distinct workdays literally.
- Check every time boundary.
- Rename equipment-sensitive stations.
- Assign the most constrained shift first.
Midweek edits and Monday reset
Scheduling state includes completed hours. Editing only future blue bars cannot undo work already performed in the current week.
Before a midweek change, record Hours Worked and future scheduled hours for every affected employee. Project total equals completed plus remaining. A full-timer who already worked thirty-six and receives twenty future hours is projected to fifty-six even if the new repeating template contains only forty. Removing future hours may stop further excess but cannot erase the completed counter. Satisfaction may remain negative until Monday resets the week.
Use transition rosters. If a permanent template changes Wednesday, create a temporary Wednesday–Sunday plan that keeps projected totals valid, then install the repeating template Monday morning. Do not toggle shared schedule repeatedly in an attempt to clear the warning; that can copy unintended assignments across days while remembered hours remain unchanged. Capture the current week before large edits, especially after changing opening hours or returning employees from training.
A release-day developer reply for 1.0 reconfirmed the 30–50 band, weekly memory and Monday reset. Reports of four ten-hour days staying red therefore need the already-worked counter before they support a bug claim. If the counter and all demand tooltips are valid after Monday on Build 3674, reproduce without mods, save screenshots, advance one day and submit F2. This separates stateful behavior from a persistent defect.
Action checklist
- Record completed and future hours.
- Use a transition roster until Monday.
- Do not expect toggles to erase history.
- Install major patterns after reset.
- F2 only after a clean post-Monday reproduction.
Auto-fill versus manual scheduling
EA 0.8 made auto-fill demand-aware and more capable, but the correct mode depends on whether speed, long shifts or economic placement is the constraint.
Standard auto-fill searches for longer assignments, can use shifts up to twelve hours, considers demands and assigns office desks more strategically. It may continue calculation in the background. Fast Auto-fill accepts shorter two-hour blocks to return quickly. Use standard mode for a stable weekly roster and Fast mode for an emergency draft, then audit fragments. Neither mode can create candidates with missing availability or decide that an unprofitable hour should be closed.
Use one-day fill when only a changed weekend or special pattern needs repair. Copy and paste a verified day, then adjust demand-specific differences. Rename workstations by role, room and index—such as Register-A-Front or Lawyer-03-East—to make filters meaningful. Disable shared schedule when days differ; developer replies repeatedly identify the day selector and shared toggle as the solution to players who unintentionally changed all days.
Manual scheduling is justified for peak overlaps, factories with recipe-stage bottlenecks, offices with thin weekend demand and employees with unusual combinations. Start from auto-fill, export or record uncovered cells, then optimize the expensive hours. Re-run the feasibility audit after every manual edit: station coverage, status bands, prohibited times, distinct days, desk compatibility and management assignment. A schedule that looks elegant but violates one critical demand is not optimized.
Action checklist
- Use standard fill for stable long shifts.
- Use Fast fill only as a draft.
- Fill or copy one day when appropriate.
- Disable shared mode for differing days.
- Audit every automated result.
Opening-hour economics
The cheapest scheduling improvement is often closing an hour whose incremental gross contribution does not cover loaded labor and operating cost.
For each edge hour, estimate contribution from sales or services minus product cost, then subtract wages, benefit allocation, security, cleaning and any role-specific expense required only because the business remains open. The hour is economically justified when incremental contribution is positive and its staffing does not create an expensive residual elsewhere. A profitable hour can still be strategically poor if covering it forces a new ten-person insurance minimum or another HR manager.
Test edges in complete comparable blocks. Close 20:00–22:00 for one normal week, then reopen for another while holding price, stock, marketing and staffing skill stable. Avoid hype, shortage and rival attack periods. Compare total business contribution, not revenue. Weekend and weekday behavior differ; test them separately. Do not copy an unversioned website's recommended hours because EA 0.5 reduced off-hour traffic and later business balancing changed demand.
Use break-even customers per hour: loaded incremental hourly cost divided by average contribution per completed sale or service. If opening requires one cashier at thirty loaded dollars and average contribution is ten, at least three otherwise-lost sales are required. Customers who would have arrived earlier after a schedule change are not necessarily incremental, so examine weekly totals. The goal is a stable profitable calendar, not maximum blue area.
Action checklist
- Calculate edge-hour contribution.
- Include threshold benefit and management costs.
- Test weekdays and weekends separately.
- Reject static recommended-hour tables.
- Optimize profit, not open duration.
Cleaning, security and support shifts
Support labor should be scheduled from the risk or condition it controls. Copying cashier coverage wastes payroll and can still miss the moment support is needed.
Schedule cleaning from observed dirt generation and the cleanliness threshold that affects customer satisfaction. One concentrated shift may restore the business more efficiently than continuous coverage, but large or high-traffic stores may need overlap. Inspect cleanliness at several times per day for a week, note when it declines, and place cleaner hours before the harmful interval. Part-time workers are often ideal because the required weekly total fits 10–30 hours.
Schedule security from theft evidence and asset exposure. Measure inventory loss or security insight with and without staffed equipment across comparable periods. Loaded security cost should be below avoided loss plus any protected customer effect. A guard present during empty hours adds no value merely because the store is open. Different products and neighborhoods produce different risk, so do not transfer one jewelry roster to a low-value food shop.
Warehouse drivers and HQ managers are support roles with downstream multipliers. Their schedule must overlap the plan's execution requirements, compatible vehicle and active business state. A missing route can stock out several stores even when each retail roster is perfect. Track failed or delayed deliveries, plan status and destination stock. Prioritize support schedules by contribution at risk across all dependent businesses rather than the worker's own location revenue.
Action checklist
- Observe dirt by time before scheduling cleaners.
- Measure theft before adding security hours.
- Use part-time staff for compact support needs.
- Verify HQ and warehouse plan execution windows.
- Rank support gaps by downstream contribution at risk.
Sickness, training and replacement transitions
A roster needs state transitions for absence, manual training, passive development and permanent replacement. Static weekly coverage is only the baseline.
An employee assigned to an HR manager can receive temporary sick coverage. Confirm the worker is in the HR group and inspect the store rather than assuming the sick icon means an empty station. Temporary coverage incurs cost that should be measured in EconoView. Without HR, maintain a spare compatible worker, accept closure risk, or shorten hours. Redundancy is most valuable for stations whose absence closes the business or disrupts many dependents.
Manual training can unassign the employee and remove them from the schedule. Before sending a batch, save business, workstation, template, HR group and uniform in a return list. Train only as many workers as the remaining roster can cover, or stagger by shift. HR-managed passive training avoids unassignment and is preferable for operational continuity, though it develops more slowly and consumes management capacity and training cost.
Permanent resignation, retirement or poaching is a headhunter responsibility when linked through the employee's HR manager. A replacement may arrive after delay and should be audited for schedule and demands. Build transition templates that identify which flexible workers can cover each critical vacancy. Never rely on a circular chain where the only headhunter able to replace HQ staff has no independent successor.
Action checklist
- Verify HR sick-cover assignment.
- Keep a return list for manual training.
- Stagger training by coverage block.
- Link HR groups to permanent replacement.
- Map backup workers for critical stations.
Schedule cost and utilization metrics
A roster should be evaluated with coverage, compliance, cost and output metrics. Any one alone can reward the wrong behavior.
Coverage rate equals staffed required station-hours divided by required station-hours. Demand-compliance rate equals employees with a valid critical band and hard schedule constraints divided by scheduled employees. Loaded labor cost includes wages, insurance allocation, temporary cover and amortized recruiting or training. Utilization compares productive transactions or units with staffed station-hours. Track all four weekly and by role.
Peak capture measures incremental contribution during intentionally overlapped hours. Edge-hour margin measures contribution after incremental cost for first and last opening blocks. Schedule volatility counts changed employee-hours after Monday; high volatility predicts remembered-hour confusion and satisfaction problems. Vacancy exposure sums contribution at risk in cells with only one compatible employee and no HR cover. These metrics turn vague roster anxiety into a prioritized repair list.
Use thresholds as local controls, not game rules. Aim for 100 percent coverage of mandatory profitable cells, zero critical demand violations, and management groups below about eighty-five percent of capacity. Investigate utilization that declines for two comparable weeks. Flag any midweek change projecting a worker above the status maximum. A high coverage rate with negative edge margins means opening-hour design—not recruiting—is the failure.
Action checklist
- Track coverage and demand compliance.
- Calculate loaded cost by role.
- Measure output per staffed hour.
- Count midweek schedule volatility.
- Separate undercoverage from unprofitable coverage.
Diagnostics and Monday audit
Most schedule failures can be isolated by following the same order: business day, opening hours, station, employee compatibility, remembered hours, demands and management state.
If the store says open but unstaffed, select the correct weekday, inspect the opening bar, then every required workstation row. Confirm shared schedule did not overwrite the day. Search for employees unassigned by training or business-type change. If enough names exist but auto-fill leaves gaps, compare each worker's demands with the uncovered time. An impossible candidate pool is not an algorithm bug.
If full-time or part-time remains red, add completed Hours Worked to future hours. Check Monday reset, exact tooltip boundary, distinct workdays and prohibited intervals. If satisfaction still falls, inspect non-schedule demands; two Important omissions can outweigh a met Critical status. Do not add random hours to make the checkbox change, because those hours may be remembered and deepen the problem.
Every Monday after reset, audit workers below or above target bands, uncovered profitable cells, low-utilization overlaps, employees completing training, sickness clusters, expiring candidates, HR groups above buffer and replacements not yet seated. After Build 3674 update, advance one normal day before concluding training or schedules remain frozen. For a clean persistent defect, capture build, mods, difficulty, before/after schedule and employee tooltip, then submit F2.
Action checklist
- Diagnose in a fixed order.
- Add completed and future hours.
- Inspect all demands, not status alone.
- Run the audit after Monday reset.
- Submit reproducible Build 3674 evidence.
Scheduling record and reproducible tests
A small data record supports profitable rosters and distinguishes current mechanics from old community habits.
Employee fields should include role, wage, status, target hours, completed hours, future hours, distinct days, prohibited times, workstation demands, assigned business, HR owner, training state and satisfaction trend. Shift fields should include business, day, start, end, role, station, required count, assigned employee, loaded cost and observed output. Record why a cell exists: base operation, measured peak, cleaning threshold, theft control or contingency.
To test a status boundary, begin Monday with one employee, no prior hours and an unchanged schedule. Observe 29, 30 and 31-hour variants across separate weeks, capturing the tooltip and result after midnight. To test a time restriction, move only one block across the tooltip boundary. To test auto-fill, preserve the same employees and opening matrix, run standard and Fast modes in copied saves, then compare uncovered cells, fragments, demand violations and loaded cost.
To test opening hours, alternate matched full weeks and control stock, prices, marketing, skill, rivals and market events. To test overlap, add one station only during named hours and measure incremental contribution. Publish the Build 3674 label, settings and raw schedule with any conclusion. Retest when patch notes mention scheduling, demands, business hours, customer flow or employee state; never upgrade a local observation into a permanent cap without that boundary.
Action checklist
- Keep employee and shift tables separate.
- Record the business reason for each cell.
- Start boundary tests on Monday.
- Change one schedule variable at a time.
- Publish controls with conclusions.
Transition arithmetic and exception ownership
A schedule change during the week needs both an hours calculation and an operational bridge. Treating a new repeating template as if the week started again is the main source of false status errors.
For every affected employee, calculate projected current-week hours = completed hours + unchanged future hours + newly assigned future hours − removed future hours. Example: an employee has completed twenty-four hours by Wednesday, retains sixteen future hours, gains twelve and loses eight. The projection is 24+16+12−8=44, which remains inside the full-time band. The repeating template might display forty, but that number cannot prove the current week is valid because already-completed cells remain in memory until Monday.
Build a transition roster for the remaining days and install the permanent pattern after the Monday reset. Mark each temporarily uncovered or overlapped cell with an owner, deadline and bridge. A Friday vacancy for Saturday evening might be owned by the store manager, bridged by a compatible part-timer and scheduled for a recruitment review Monday. Close the exception only when the intended worker is seated and the next comparable interval confirms coverage. Candidate acceptance and a green notification are intermediate states, not proof of productive placement.
Price resilience explicitly. If a fourteen-hour part-timer costs $420 weekly including wages, insurance allocation and recruiting amortization but prevents an expected six hours of closure at $180 contribution per hour, the value protected is $1,080 and net risk-adjusted benefit is $660. If HR temporary coverage reduces expected closure to one hour, the same reserve becomes uneconomic. Replace illustrative inputs with observed margins, and keep the assumptions beside the roster so a later demand or benefit change triggers recalculation.
Action checklist
- Calculate completed plus adjusted future hours.
- Use a remaining-week transition roster.
- Give every exception an owner and deadline.
- Verify productive placement after notifications.
- Compare reserve cost with expected vacancy loss.
| Transition field | Worked value |
|---|---|
| Completed by Wednesday | 24 |
| Unchanged future | 16 |
| New hours | 12 |
| Removed hours | 8 |
| Projected total | 44 |
| Permanent change date | Monday after reset |
Roster reconciliation and decision thresholds
A complete audit reconciles station-hours, employee-hours, overlap and paid nonproductive time. The totals reveal gaps that a visually tidy calendar can hide.
Record total required station-hours, covered required hours, deliberately uncovered unprofitable hours, justified overlap and paid nonproductive time. The identity assigned employee-hours = covered required hours + justified overlap + paid nonproductive time should explain every scheduled hour. Separately, required station-hours = covered hours + deliberately uncovered hours + accidental uncovered hours. Accidental gaps are failures; deliberate gaps need an edge-hour margin note. If the identities do not balance, locate the unmatched day, role and workstation before judging labor cost.
Compare Standard and Fast auto-fill with the same employee pool and opening matrix. Count uncovered profitable cells, two-hour fragments, Critical demand violations, workers outside status bands, loaded wage-hours and single-point cells. Reject any Critical violation before comparing cost. Among feasible rosters, select the lowest loaded cost unless a slightly more expensive plan materially reduces vacancy exposure. This preserves the scheduler's strength as a constraint solver while keeping economic judgment outside the algorithm.
For opening or overlap decisions, use interval contribution = incremental transactions or units × gross contribution per unit − loaded interval labor − incremental operating cost. Investigate an ordinary edge after two comparable negative weeks, but act sooner on a consistently empty staffed cell. Add a second station only when measured incremental contribution clears its loaded cost plus a risk margin. Store screenshots of opening bars, workstation rows and employee tooltips with the numbers; a weekly total cannot prove hours occur on the right day or compatible station.
Action checklist
- Reconcile required, covered, overlap and paid time.
- Identify accidental gaps separately.
- Compare auto-fill modes on identical inputs.
- Reject Critical violations before cost ranking.
- Use interval contribution for edge and overlap decisions.
| Metric | Formula | Action threshold |
|---|---|---|
| Coverage rate | Covered required / required | Repair profitable accidental gaps |
| Demand compliance | Valid scheduled workers / scheduled workers | Reject roster with Critical violations |
| Edge margin | Incremental contribution − loaded edge cost | Close after controlled negative evidence |
| Vacancy exposure | Single-point hours × contribution per hour | Add bridge when expected loss exceeds reserve cost |
| Reconciliation difference | Assigned − covered − overlap − nonproductive | Must equal zero |
Research ledger
These links establish mechanics or provide a reproducible lead. Any balance-sensitive number still has to be checked in the current save.
Questions answered
What counts as full-time and part-time?
Current developer and F1-linked evidence uses 30–50 remembered weekly hours for full-time and 10–30 for part-time. Read the live tooltip at the shared thirty-hour boundary and use 31 or 29 when a robust schedule matters.
Why is forty hours still marked wrong?
The employee may have completed additional hours earlier in the same week. Add Hours Worked to all future scheduled hours. The memory resets Monday morning; changing blue bars midweek does not erase completed work.
Can auto-fill optimize profit?
No. It improved substantially in EA 0.8 and considers demands, but it fills requested coverage. It does not know whether a late hour, second register or fully occupied office is economically justified. Audit output, loaded labor, vacancy exposure, retention risk and edge-hour contribution.
What is the maximum automatic shift length?
EA 0.8 Build 3014 states that auto-fill can schedule shifts up to twelve hours. This is an auto-fill capability, not proof that every employee demand combination accepts every twelve-hour pattern.
How many workers cover a 24-hour seven-day station?
The station needs 168 hours. Four full-timers at forty-two hours can cover it arithmetically, such as four six-hour daily shifts, if demands permit seven days. Mixed rosters may be cheaper or more compatible. Add redundancy based on HR coverage and vacancy risk.
Should every installed workstation be staffed?
No. Staff stations justified by customer demand, queues, production bottlenecks or risk control. Empty capacity is cheaper than a worker whose incremental output does not cover loaded labor cost across a comparable operating week. Confirm the conclusion with the business insight panel and matched intervals: a queue can justify overlap, while an idle installed station does not create demand. Preserve stock, price, marketing and opening hours during the comparison.
How should I compare two feasible rosters?
Reject any roster with a Critical demand violation, then compare loaded labor cost, profitable coverage, fragments, justified overlap and vacancy exposure. Use the same opening matrix and employee pool. A slightly more expensive roster can be preferable when it removes a single point of failure whose expected closure loss exceeds the added wages and benefits. Record the contribution assumption and review actual results after a comparable week; do not choose only by the number of filled blue cells.
Does HR auto-fill schedules?
The scheduler provides auto-fill capabilities associated with the management workflow, and EA 0.8 made them demand-aware. HR's separate core functions include passive training and temporary sick coverage. Always inspect the generated schedule rather than assuming it is profit-optimal. Compare required station-hours with assigned compatible hours, reject Critical demand violations, identify short fragments, price every overlap and verify each headquarters plan after the roster is installed. Standard mode can search for longer feasible blocks, while Fast mode is better treated as an emergency draft. Neither mode decides whether an opening interval earns enough contribution to remain open, and neither guarantees resilience when one uniquely compatible worker becomes unavailable. Review actual queues, transactions, production and vacancy exposure during comparable weeks before adopting the generated pattern as the permanent schedule.
When should I edit a permanent schedule?
Monday immediately after the weekly reset is safest. For a midweek change, build a temporary transition roster using completed plus future hours, assign an owner and bridge to every gap, then install the permanent repeating pattern after reset. Verify coverage during the next comparable interval before closing the exception.