Shift Design

Shift Design is the practice of building the catalog of shifts that Schedule Generation draws from. The shift catalog is the input to optimization — its quality bounds the quality of every schedule built afterward. A bad shift catalog produces bad schedules no matter how good the optimizer is.

For workforce management, shift design is foundational and underrated. Most WFM teams inherit a shift catalog and rarely revisit it; revisiting it is one of the highest-leverage interventions available because every weekly schedule is constrained by the catalog. This page documents the practice.

What a shift is

A shift is a structured time block specifying:

• Start time — when the agent's day begins
• Duration — total length (e.g., 8.5 hours including breaks)
• Days of week — which days the shift applies (Monday-Friday, Tuesday-Saturday, weekend-only, etc.)
• Break placement — when meal and rest breaks fall within the shift
• Off-phone activity slots — coaching windows, training time, administrative time
• Skill requirements — for multi-skill operations, which skills the shift requires the agent to staff

A shift catalog is the collection of all valid shifts an organization offers. Schedule generation assigns agents to shifts from this catalog.

Why shift design matters

Three reasons the catalog is high-leverage:

1. The catalog bounds the achievable coverage curve. If demand peaks at 11:00 AM but no shift starts before 9:00 AM, you cannot match the peak with a single-shift assignment — the optimizer must layer overlapping shifts to approximate the curve, increasing total FTE.
2. Bad catalogs amplify schedule cost. Mismatched start times, awkward break placement, or insufficient shift variety force the optimizer into expensive workarounds (split shifts, frequent overtime, undesirable shift premiums).
3. The catalog encodes labor agreements and constraints. Union rules, contract terms, regulatory limits all live in shift definitions. Schedule quality starts with whether the catalog correctly represents what's allowed.

The forecasting accuracy → schedule quality story is well-known. The shift catalog → schedule quality story is less talked about but matters at least as much.

Designing a shift catalog

The basic procedure:

1. Plot the demand curve. From the forecast, compute the FTE requirement per interval across a representative week. The curve shows where coverage is needed.
2. Identify coverage targets. For each interval, the FTE requirement is the target. Sum across the day to get daily FTE-hours.
3. Generate candidate shifts. Build a catalog of plausible shifts respecting labor constraints (minimum / maximum durations, allowed start times, break rules, regulatory requirements).
4. Evaluate fit. For each candidate shift catalog, run a schedule-generation pass against the forecast to see how well the catalog supports the demand curve.
5. Iterate. Refine — add shifts to cover gaps, remove shifts that go unused, adjust start times to better match peaks.

Modern WFM software supports shift design optimization explicitly: a separate optimization that selects which shifts to include in the catalog given the demand curve. This is a different problem from schedule generation; the inputs and constraints are different.

Coverage curves and shift profiles

A practitioner's mental model: the demand curve is a continuous function of time; each shift is a step function (1 during the shift, 0 outside it); the schedule is a sum of step functions; the goal is to match the demand curve with the sum.

Mathematically, with $D(t)$ as demand at time $t$ and $n_s$ as the number of agents on shift $s$:

$\text{Coverage}(t)=\sum _s{n}_s\cdot \mathbb{𝟙}[t\in s]$

The shift catalog determines which step functions are available. A catalog with shifts only at 9:00 AM and 5:00 PM forces the coverage curve to look like two boxes; a catalog with shifts every 30 minutes from 6:00 AM to 12:00 PM allows the coverage curve to track demand much more closely.

The cost: shift variety. More shifts → finer coverage tracking → higher administrative overhead and harder optimization.

Common shift design parameters

Practical decisions a shift catalog makes:

Start times

How granular? Common choices:

• Hour boundaries (6:00, 7:00, 8:00 AM) — simplest; coarse coverage
• 30-minute boundaries (6:00, 6:30, 7:00 AM) — typical for contact centers; reasonable balance
• 15-minute boundaries — finer; useful for highly variable demand
• Continuous (any minute) — only used in advanced flexible models

The granularity choice trades coverage precision for catalog complexity.

Durations

• Fixed duration (all 8-hour shifts) — simplest; uniform agent treatment
• Multiple durations (4-hour, 6-hour, 8-hour, 10-hour) — supports both full-time and part-time pools and lets the optimizer pick the right length per coverage need

Break placement

• Anchored (break must occur 4 hours into the shift) — strict
• Window-based (break must fall within hours 3.5-4.5 of the shift) — more flexible; standard in most modern WFM
• Optimizer-chosen (the schedule optimizer places breaks to maximize coverage) — most flexible; complicates the optimization

Off-phone activity slots

Reserved time within shifts for coaching, training, administrative work. The catalog can:

• Pre-allocate fixed slots (e.g., 30 minutes coaching every Tuesday at 2:00 PM)
• Pool the time as a budget the agent's supervisor allocates dynamically — better fit for Variance Harvesting practice

Days of week

• Monday-Friday shifts vs. weekend shifts vs. rotating patterns
• Days-of-week patterns are often where labor agreements have the most variation

Common WFM pitfalls

• Inheriting a stale catalog. The most common failure mode. Catalogs designed for an old demand pattern persist after the demand pattern shifts; the optimizer compensates with workarounds; nobody notices the catalog is the constraint.
• Too few shifts. A catalog with only 3-4 shifts forces overstaffing during demand troughs because no shorter shift is available.
• Too many shifts. A catalog with 50+ shifts produces administrative overhead, agent confusion, and harder schedule optimization without proportional coverage gain.
• Break placement that fights coverage. Anchored breaks at fixed times (e.g., everyone breaks at 12:00 PM) creates an artificial coverage hole exactly when it most often hurts.
• Not modeling regulatory constraints. Maximum daily hours, mandatory rest periods, weekend rotation rules — if these aren't in the catalog, the optimizer will produce schedules that violate them.
• Treating off-phone time as schedule waste. Off-phone time is not waste; it's coaching, training, and the substrate for Variance Harvesting practice. Catalogs should preserve enough off-phone slots to support those activities.

Shift catalog refresh cadence

Most organizations refresh their shift catalog far less often than demand changes. A practical cadence:

• Quarterly review — does the catalog still match demand patterns? Are unused shifts persisting? Are coverage gaps recurring at specific intervals?
• Annual full redesign — re-evaluate the catalog from scratch given the current demand pattern, agent pool, and labor environment
• Triggered refresh — when major demand structure changes (new product launch, vendor migration, channel shift), redesign immediately rather than waiting for the calendar review

The practical reality: most organizations skip these reviews. Triggering them is one of the lower-cost, higher-impact moves available to a WFM team.

Connection to forecasting

Shift design's primary input is the demand forecast. The forecast accuracy bounds shift design quality: a noisy forecast produces a catalog optimized for noise. Best practice: design the catalog against a multi-week aggregate forecast (smoothing the noise), then validate against the realized variance.

Probabilistic Forecasting is particularly useful here. Designing the catalog against the P50 forecast produces a catalog that's right on average but fragile under variance; designing against the P80 forecast (with appropriate cost trade-off) produces a more resilient catalog.

Maturity Model Position

In the WFM Labs Maturity Model™, shift design is foundational at Level 2 but takes on dramatically different character at higher levels. The catalog itself is a maturity artifact: how often it is refreshed, how it accommodates off-phone time, and whether it is designed against distributions all signal where an organization sits.

• Level 1 — Initial (Emerging Operations) — shift catalogs are informal or inherited; demand-curve fit is not analyzed.
• Level 2 — Foundational (Traditional WFM Excellence) — a fixed catalog is in place but rarely refreshed; designed against point forecasts; off-phone time is treated as schedule waste; break placement often fights coverage.
• Level 3 — Progressive (Breaking the Monolith) — quarterly catalog reviews are routine; off-phone time is pooled as a budget for supervisor-allocated coaching and learning (the substrate for Variance Harvesting); window-based break placement replaces anchored breaks; designed against P50/P80 to be more resilient under variance.
• Level 4 — Advanced (The Ecosystem Emerges) — flexible catalogs with finer start-time granularity; catalog optimization runs alongside schedule optimization; catalog respects distributional demand rather than point estimates; off-phone slots are designed for self-allocation.
• Level 5 — Pioneering (Enterprise-Wide Intelligence) — continuous-start-time / on-demand shift composition replaces static catalogs; agent-driven shift selection is the dominant pattern; the catalog distinction blurs into a flexible workforce model.

Shift design is a legacy-to-progressive practice — its presence at L2 is correct, but the stale catalog is one of the highest-leverage and most overlooked maturity lifts: refreshing it and pooling off-phone time alone often moves an organization meaningfully into Level 3 territory.

References

• Koole, G. "Workforce scheduling." Call Center Optimization, chapter on scheduling. cs.vu.nl/~koole/ccmath/book.pdf.
• Cleveland, B. Call Center Management on Fast Forward (3rd ed.). ICMI Press, 2012. Practitioner discussion of shift design.
• Gans, N., Koole, G., & Mandelbaum, A. "Telephone call centers: tutorial, review, and research prospects." Manufacturing & Service Operations Management 5(2), 2003.

See Also

• Scheduling Methods — overview where shift design sits in the workflow
• Schedule Generation — the optimization that consumes the shift catalog
• Adherence and Conformance — the measurement layer
• Forecasting Methods — the demand forecast that drives shift design
• Probabilistic Forecasting — designing catalogs against forecast distributions
• Variance Harvesting — operational practice that depends on off-phone slot availability
• WFM Processes — broader operating model
• WFM Roles — Scheduler role
• Multi-Objective Optimization in Contact Center — when shift design has multiple objectives