Decisions
Workforce forecasting: roster against a path, not a hunch
Nurses, agents, cooks, field techs — the cost is people on a clock. A staffing forecast is arrivals or contacts by interval, with a band wide enough for a call-out list, not a single headcount that makes Tuesday look like Saturday.
Updated Aug 26, 2026·8 min read
The interval is the product
Nurses, agents, cooks, field techs — the cost is people on a clock. A staffing forecast is arrivals or contacts by interval, with a band wide enough for a call-out list. A daily headcount that makes Tuesday look like Saturday is how you burn overtime and still miss the surge.
15-minute load, not a daily headcount
Contact centers live in 15- or 30-minute queues. EDs live in hours. Kitchens live in the lunch wave. Forecast the grain you schedule. You can roll 15-minute paths up. You cannot honestly split a day into quarters after the fact.
Shrinkage is not demand
PTO, training, and a 10% off-phone factor belong in the rostering tool. If you mix them into the demand series, you teach the model your absence policy. Forecast the work. Apply shrinkage where humans are assigned.
- Separate volume (contacts, patients, tickets) from handle time if both move.
- Holidays and paydays are calendar, not a mysterious Monday effect.
- A nowcast of today’s arrivals can still save the evening roster after the official file is late.
Roster the median, list the p90
Staffing to the mean guarantees you are short on the fat days. Staffing to the max every interval is how you go broke. A quantile is the product: base roster on p50, keep an on-call list for p90. Static “high volume” thresholds from 2023 will page every weekend and miss a weird Wednesday.
A goal on contacts, not on “busy”
Push arrivals or offered contacts as a signal. Predict.ai will pick up weekday shape and holidays, race models on that interval, and keep the path live. Alerts fire when actuals leave the band — you add a closer because Wednesday broke, not because a threshold said “high.”
Driver discovery will look for campaigns, outages, weather, a product launch. Shrinkage stays in WFM. The analyst can explain a spike without waiting on a weekly packet. That is the gap with a spreadsheet that still thinks last year’s interval mix is a law.
FAQ
- What grain should a contact center forecast use?
- The same interval you schedule: 15 or 30 minutes, by queue. Daily totals cannot place people on a shift. You can roll 15-minute forecasts up. You cannot honestly split a day into quarters after the fact.
- How do I include shrinkage?
- Forecast the work (contacts, patients, tickets). Apply shrinkage in the rostering tool. Mixing PTO into the demand series teaches the model your absence policy, not the customers.
Keep going
Guide
Nowcasting
Nowcasting is a forecast with a horizon of this afternoon. The official close has not posted; you still have to staff the evening or dispatch the next hour. You use faster series — traffic, partial files, a weather nowcast — to estimate the present.
Guide
Prediction intervals
A prediction interval is a range the future value is expected to fall in, not a promise. p50 is the middle path; p10–p90 is a band for planning stock, staff, and cash without pretending the future is a single line.
Guide
Holidays and calendar effects
A calendar is the cheapest accurate feature in the building. Christmas, payday, school terms, a 4-day week, Ramadan, Chinese New Year — if the model has to rediscover those from residuals, you are wasting folds on a date table.
Use case
Healthcare staffing forecasting
Translate expected patient workload into staffing needs by unit, role, shift, and skill.
Use case
Patient demand forecasting
Anticipate arrivals, admissions, bed demand, and procedure volume across facilities and service lines.