WorkAxle Demand Forecasting & Roster Optimization is a workforce management module that does two jobs: it predicts how many people you need, then builds the roster to match. Forecasting reads your demand history and seasonality, tests multiple time-series models in parallel, and your team selects the best-fit model for each location. Roster optimization turns that forecast into balanced, fair, compliant seasonal rosters and re-optimizes them as inputs firm up.
Staff for the week
that's actually coming.
Without forecasting, every schedule is a guess, and every guess shows up on the labor line. WorkAxle reads your demand data, runs several prediction models in parallel, and selects the model that fits each location. Then it builds balanced, fair rosters from that forecast and re-optimizes them as the season firms up.
Overstaffed Monday. Understaffed Friday. Every week.
When every schedule is a guess.
Copy last week. Paste last week's schedule and hope demand holds. It usually doesn't. You end up paying people to stand around on slow days and scrambling for cover on the peaks.
Gut feel. An experienced scheduler builds from memory. That works right up until she leaves, or until a holiday, a storm, or a promotion rewrites the pattern she had in her head.
Both habits look backward. They build next week's schedule from a record of what already happened, with no read on what demand is about to do.
When the schedule is data-driven.
Historical demand data and custom business drivers flow into the forecasting engine. Multiple models run against your actual data. The model your team selects drives shift generation, headcount recommendations, and labor cost projections.
You don't guess how many people you need Friday. The system tells you, at the granularity of your data, down to 15-minute intervals if that's what you feed it.
→ That's the forecasting engine. Every section below explains how.
From raw data to staffed schedule, automatically.
The engine ingests your demand data.
Historical demand at your data's granularity. Holidays and time features are handled for you.
It tests multiple models per dataset.
The engine runs multiple time-series models in parallel against your data; your team selects the best-fit model per location.
Every run sharpens the read.
Forecasts refresh as new actuals arrive, and every run lets our customers get tighter with their operations.
Forecasts feed into scheduling.
The chosen forecast drives shift generation, headcount recommendations, and labor cost projections inside the same platform, so the numbers move into scheduling instead of being rebuilt in a separate spreadsheet.
Roster & Schedule Optimization
The forecast tells you how many people you need. The solver turns that into the roster itself. Building a season of fair, compliant schedules for a unionized workforce is an operational-research problem, and WorkAxle treats it like one.
Balanced and fair by construction.
At the start of a season, the solver builds rosters for unionized environments that are balanced from day one: a sound full-time and part-time mix, hours distributed evenly across the team, no favoritism.
Compliant inside the solve.
Union working rules, breaks, rest periods, maximum shifts, and shift lengths are honored inside the optimization, not patched afterward. The roster hits your throughput and service-level targets while protecting the margin.
Re-optimized as the season firms up.
As real inputs replace estimates, the solver re-optimizes within the existing shift structure. In a union setting you cannot rip people off their shifts, and the engine respects that.
You configure it from the front end.
Penalty points, ratios, throughput targets, who the optimization applies to: your team sets all of it in the interface. Operators can re-run the solver with adjusted parameters, tweaking a penalty weight or a constraint threshold to balance service quality and labor cost. No code, no vendor ticket.
The output lands in your scheduler.
The solver returns recommended shift templates and work-cycle templates that flow straight into scheduling. Your team and your unions review before anything publishes.
→ Day-of changes, swaps, no-shows, and replacements live in Scheduling & Rostering.
What the forecasting engine does.
| Capability | What it does |
|---|---|
| Flexible demand signals | Forecasts from your historical demand data at the granularity your data supports. |
| Multi-model comparison | Runs multiple time-series models in parallel per dataset to fit the best model per location, no data science team required. |
| Sharper with every run | Forecasts refresh as new actuals arrive; every run lets our customers get tighter with their operations. |
| Granular interval forecasting | Supports 15-minute, 30-minute, hourly, or daily intervals matching your ingested data. |
| Direct scheduling integration | Forecasts flow into shift generation and labor optimization inside the same platform, instead of living in a separate spreadsheet. |
| Multi-location support | Separate forecasts per location, each with its own demand drivers and model configuration. |
Same engine, different demand drivers.
Security & Guarding
Contract-driven demandPost coverage, event staffing, site minimums. Forecasting meets contractual obligations without over-deploying guards, which protects the margin on fixed-fee contracts.
See solutions →Public Sector
Budget-cycle demandService levels set by mandate, headcount set by budget. Forecasting shows what coverage each budget scenario actually buys before the fiscal year locks.
See solutions →Manufacturing & Processing
Production-line demandLine schedules and production targets drive headcount. Forecasting converts the production plan into crew requirements per line, per shift, before the week is built.
See solutions →Airport Operations
Flight-schedule demandPassenger volume forecast at terminal level in 15-minute intervals. The solver builds rosters that meet service-level priorities while holding to collective-agreement constraints.
See solutions →Retail
Sales-pattern demandHistorical sales patterns predict headcount at your data's granularity, so you avoid the overstaffed lull at 10 AM and the gap at the lunch rush.
See solutions →Healthcare
Patient-volume demandPatient volume and acuity patterns set the baseline. Forecasting projects coverage ahead of the roster, so minimum staffing is planned in advance.
See solutions →Construction
Project-phase demandProject phases set crew demand. Forecasting maps trades and headcount to the build schedule, so crews land on site when the phase actually starts.
See solutions →Port Operations
Vessel-schedule demandVessel arrivals drive labor demand in bursts. Forecasting converts berth schedules into gang requirements days ahead of dispatch.
See solutions →Forecasting and rostering, after WorkAxle.
Your questions, answered.
How does WorkAxle forecast labor demand?
The engine ingests your historical demand data. It runs multiple time-series models in parallel per dataset; your team selects the best fit, and it refreshes as new actuals arrive. Every run lets our customers get tighter with their operations. Forecasts match your data granularity, anywhere from 15-minute to daily intervals.
What data sources does WorkAxle use for demand forecasting?
Historical demand data, with holidays and time features handled for you, at the granularity your data supports.
How is WorkAxle's forecasting different from other WFM platforms?
WorkAxle runs multiple models in parallel; your team selects the best fit per location. For complex operations, a mathematical optimization solver generates compliant seasonal rosters and re-optimizes them as the season firms up.
Can WorkAxle forecast demand for multi-location operations?
Yes. Each location gets its own forecast, with its own demand drivers, and your team selects the model per location. WorkAxle's multi-location forecasting is built for operations that run many sites under one plan, from retail chains to complex airport operations.
Does the forecast connect to scheduling automatically?
Yes. Forecasts flow into shift generation, headcount recommendations, and labor cost projections inside the same platform, instead of being rebuilt in a separate spreadsheet.
Can WorkAxle build fair rosters for unionized teams?
Yes. The optimization solver builds balanced seasonal rosters: full-time and part-time mix, evenly distributed hours, no favoritism. As the season approaches it re-optimizes within the existing shift structure, and penalty points, ratios, throughput, and who it applies to are all configured from the front end.
See what data-driven scheduling actually looks like.
What is the Compliance Stack?
Read the architectural foundation WorkAxle is built on, from the rule engine up.
How does WorkAxle compare?
See why organizations switch from legacy WFM platforms to WorkAxle, side by side.
Ready to see forecasting in action?
Bring your demand data. We'll run a live multi-model comparison on it.