Store Labour Scheduling

Weekly shift assignment · Demand-driven

Assign store associates to shifts across a planning week so that hourly forecast traffic is covered at minimum labour cost, subject to worker availability, weekly hour caps, and fair-scheduling laws (e.g., NYC Fair Workweek, San Francisco Formula Retail Employee Rights). Integer programming with a staffing-pattern formulation; greedy + local-search heuristic in the solver. Connects empirical labour-and-sales evidence (Kesavan & Terwiesch 2015) to operational shift design.

Why it matters

Labour is retail’s largest controllable cost — and a top driver of conversion

~13%
Typical store-labour share of retail revenue — the largest controllable cost line at most chains.
Source: NRF retail expense benchmarks.
+10%
Conversion-rate lift documented when staffing is matched to traffic vs. flat shifts (Kesavan-Terwiesch 2015 grocery study).
Source: Kesavan & Terwiesch (2015), Production & Operations Management.
14 days
Advance-notice requirement under NYC Fair Workweek and similar laws — a hard constraint that forces deterministic schedules.
Source: NYC Department of Consumer & Worker Protection Fair Workweek Law.
~$0.5B
Estimated U.S. retail wage-theft penalty exposure from non-compliant scheduling — misalignment between intent and execution is costly.
Source: National Employment Law Project; industry settlements.

Where the decision sits

Operational, weekly · downstream of forecast, upstream of execution

Labour scheduling sits between traffic forecasting (an upstream prediction problem) and shift execution (the operational reality of swaps, no-shows, and overtime). At a typical chain this is run weekly per store, optimised across ~10-50 associates and ~84 hourly slots (12-hour day × 7 days). Production systems integrate with workforce-management software (Kronos, Workday); the solver below illustrates the underlying MIP. Compare with scheduling family formulations; the structural cousins are healthcare nurse rostering and call-centre staffing.

Forecasthourly traffic
Schedule shiftsweekly MIP
Publish14-day notice
Executeswaps, no-shows

Problem & formulation

Worker-shift MIP with demand coverage and fairness constraints

OR family
Integer Programming (rostering)
Complexity
NP-hard (rostering generic)
Solver realism
★★ Greedy + 1-opt swap
Reference
Pinedo (2016); Kesavan-Terwiesch (2015)

Sets and parameters

SymbolMeaningUnit
\(w \in \mathcal{W}\)Worker (associate)finite
\(t \in \mathcal{T}\)Time slot (hour-of-week)\(\{0, \ldots, 83\}\)
\(s \in \mathcal{S}\)Allowed shift pattern (start-hour + length combo)finite
\(d_t\)Forecast demand (associates needed) in slot \(t\)FTEs
\(c_w\)Hourly wage of worker \(w\)$ / hr
\(a_{w,s}\)1 if worker \(w\) is available for shift \(s\)binary
\(\bar H_w\)Maximum weekly hours for \(w\)hours
\(\underline H_w\)Minimum guaranteed hours (per local law)hours

Decision variable

SymbolMeaningDomain
\(x_{w,s}\)1 if worker \(w\) is assigned to shift \(s\)binary

Objective

Minimise total wage cost across the week:

$$\min \; \sum_{w \in \mathcal{W}} \sum_{s \in \mathcal{S}} c_w \cdot \mathrm{hours}(s) \cdot x_{w,s}$$

Constraints

$$\sum_{w \in \mathcal{W}} \sum_{s : t \in s} x_{w,s} \;\geq\; d_t \qquad \forall\, t \qquad \text{(demand coverage)}$$ $$\underline H_w \;\leq\; \sum_{s} \mathrm{hours}(s) \cdot x_{w,s} \;\leq\; \bar H_w \qquad \forall\, w \qquad \text{(weekly hour bounds)}$$ $$x_{w,s} \leq a_{w,s} \qquad \forall\, w, s \qquad \text{(availability)}$$ $$\sum_{s} x_{w,s} \leq 1 \quad \text{per day per worker} \qquad \text{(no double-shift)}$$

Fair-scheduling extensions (optional)

Modern fair-workweek laws add hard constraints:

  • Predictability premium: changes within 14 days → pay penalty (encoded as variable cost).
  • Right to rest: minimum 11 hours between shifts (clopening ban).
  • Access to hours: existing part-time workers must be offered hours before new hires — preference order encoded in objective.
  • Good-faith estimate: new hires must receive a written hours estimate.

Interactive solver

10 workers · 14 4-hour shifts (2 per day) · greedy demand-fill + 1-opt

Store labour scheduler
Mix of full-time + part-time workers, demand peaks at lunch and evening
★★ Greedy + swap (near-optimal at this scale)
FT: 40 hrs cap, $20/hr
Soft penalty for excess
—
Total weekly wage cost ($)
—
Slots fully covered
—
Total understaffed FTE-hrs
—
Total overstaffed FTE-hrs
—
Worker utilisation
—
Avg hours / worker
Worker assigned to shift Forecast demand (line) Scheduled coverage (line) Understaffed slots

Under the hood

14 shifts cover the week (2 per day × 7 days, each 4 hours: morning 09-13, evening 13-17). Workers split into FT (3 default, $20/hr, 40 hr cap) and PT (rest, $16/hr, 25 hr cap). Each worker's availability is randomly drawn (avg 70% of shifts available). Demand profile: lunch peak (slots 1, 5, 9, 13) and evening secondary peak. Solver: greedy — for each understaffed slot, fill with the cheapest available worker who hasn't hit hour cap; then 1-opt swap for cost reduction. Reports cost + coverage + utilisation.

Reading the solution

Three patterns to watch for

  • FTs anchor the peaks. Lunch and evening peaks pull in full-time associates first because they need their guaranteed hours; PTs fill the off-peak gaps.
  • Trade-off cost vs coverage. Understaffing one peak slot saves an hour of wage but loses conversion. The overstaff penalty parameter tunes how aggressively to over-cover.
  • Availability bottlenecks. If many workers are unavailable on Saturday evening, the solver pays peak FT rates on a slot that should have been PT — the canvas highlights the mismatch.

Sensitivity questions

  • Add 2 more PT workers? — coverage improves; total cost may stay flat (cheaper labour fills slots that were short).
  • Tighten FT hour cap to 32 hrs? — PT coverage jumps; total cost slightly up; better Schedule Predictability compliance.
  • Higher peak demand (6 → 8)? — understaffed hours rise; need to hire or train more workers.

Model extensions

Stochastic demand

Forecast as distribution; chance-constrained or two-stage SP variant. Trade-off coverage probability vs cost.

Multi-skill workers

Cashier, floor, stockroom skills. Generalised assignment with role-eligibility.

Fair-workweek strict mode

11-hour rest, 14-day notice, predictability premium for changes — turn the soft constraints in this solver into hard ones with predictability-pay terms.

Cross-store labour pool

Workers can cover multiple nearby stores. Bigger MIP; transportation cost added.

Joint with traffic forecast

End-to-end forecast + schedule; Kesavan-Terwiesch 2015 endogeneity (more labour → higher conversion → higher demand).

Worker preferences + utility

Soft constraints: preferred days off, max consecutive days. Multi-objective with worker satisfaction.

Healthcare scheduling sibling

Nurse rostering uses essentially the same MIP with stricter regulatory constraints.

Nurse rostering →
Workforce planning sibling

Multi-month workforce sizing decision (hire/fire) sits upstream of weekly scheduling.

Workforce planning →

Key references

Kesavan, S. & Terwiesch, C. (2015).
Doctors, drugs, and dietary rules: Frontline workers and operational constraints in retail healthcare clinics.
Production and Operations Management 24(8): 1306–1322. (and related labour-and-sales papers from the same group.)
Mani, V., Kesavan, S. & Swaminathan, J. M. (2015).
Estimating the impact of understaffing on sales and profitability in retail stores.
Production and Operations Management 24(2): 201–218. doi:10.1111/poms.12237
Pinedo, M. L. (2016, 5th ed.).
Scheduling: Theory, Algorithms, and Systems.
Springer. (Comprehensive scheduling textbook; rostering chapters.)
Ernst, A. T., Jiang, H., Krishnamoorthy, M. & Sier, D. (2004).
Staff scheduling and rostering: A review of applications, methods and models.
European Journal of Operational Research 153(1): 3–27. doi:10.1016/S0377-2217(03)00095-X
Van den Bergh, J., Beliën, J., De Bruecker, P., Demeulemeester, E. & De Boeck, L. (2013).
Personnel scheduling: A literature review.
European Journal of Operational Research 226(3): 367–385. doi:10.1016/j.ejor.2012.11.029
Kabak, Ö., Ülengin, F., Aktaş, E., Önsel, S. & Topcu, Y. I. (2008).
Efficient shift scheduling in the retail sector through two-stage optimization.
European Journal of Operational Research 184(1): 76–90.
Williams, J. & Boushey, H. (2010).
The three faces of work-family conflict: The poor, the professionals, and the missing middle.
Center for American Progress. (Policy basis of Fair Workweek laws.)
NYC Department of Consumer & Worker Protection (2017–).
Fair Workweek Law (Local Law 99 of 2017).
14-day advance notice, premium pay for changes, predictability for retail and fast-food workers.

Back to the retail domain

Store labour scheduling sits in the Place × Operational cell — weekly cadence, biggest controllable cost line, increasingly regulated.

Open Retail Landing
Educational solver · greedy heuristic on a 14-shift toy schedule · production scheduling uses commercial WFM tools (Kronos, Workday, Reflexis) with full MIP backends.