Run it with less chaos · Every business
Staff Scheduling: Building a Weekly Schedule That Matches Demand
Most small service businesses schedule staff by habit and the busiest day sets the tone, while a randomized study of retail stores found that steadier, demand-based schedules raised both sales and productivity.
At a glance
- Demand-based scheduling means setting each day's staff hours from the bookings and requests customers actually make, with a stable core schedule people can plan around and a small flexible layer for the peaks.
- Unstable schedules carry a measurable cost: a study of 1,827 hourly retail and food service workers in the ILR Review found schedule instability to be a strong and consistent predictor of turnover.
- In a randomized pilot across 28 retail stores, the stores that adopted more stable scheduling practices saw median sales rise 7% and labor productivity rise 5%.
- For an illustrative salon with seven stylists on the floor every day, the week already has the right number of hours (280 scheduled against about 278 needed); moving 32 of them from Tuesday and Wednesday to Friday and Saturday is the fix, not hiring.
- A few cities and states have predictive scheduling laws for large employers, so check the rules where you operate before building a template.
Part 1 | Start from demand, not from last month's schedule
The usual small business schedule is last month's schedule with changes for whoever asked for time off. Saturday is chaos; we need more people. The sentence is often accurate about Saturday and silent about Tuesday, where the same number of people stand around a half-empty book, and a schedule built by habit tends to spread hours evenly across days that are anything but even.
Demand data is already in the booking system. Pull four weeks of booked service hours by day and by hour, then add what the calendar does not show: the requests the front desk turned away because no one was free, which a simple tally sheet at the desk can capture in a week or two. Booked hours plus turned-away hours is the demand curve, and dividing it by a target utilization of about 85% gives the staff hours each day needs, since queueing research shows waits climbing steeply once a schedule runs fuller than that.
Part 2 | Stable schedules paid for themselves in a controlled test
The best evidence on scheduling and performance comes from a randomized pilot run with Gap Inc. and researchers from the University of California Hastings College of the Law, the University of Chicago and the University of North Carolina. Across 28 stores in the San Francisco and Chicago areas, 19 treatment stores and 9 control stores, from November 2015 to August 2016, the treatment stores added practices such as consistent start and end times, the same shifts week to week, a core team with a soft guarantee of 20 or more hours, app-based shift swapping and extra staffing where it was likely to lift sales, on top of two-week advance schedules and an end to on-call shifts across the company.
The results favored stability. Median sales in the treatment stores rose 7% and labor productivity rose 5%, with treatment stores producing $6.20 more revenue per labor hour than control stores, and the researchers found that only 30% of the week-to-week swings in payroll hours were explained by changes in store traffic, with the rest coming from internal causes such as shipment errors and last-minute promotion changes.
The turnover research points the same way. An analysis of 1,827 hourly workers in U.S. retail and food service, published in the ILR Review, found that schedule instability strongly predicted who left, and that slightly less than half of the link ran through lower job satisfaction and about a quarter through conflict between work and family. A salon or clinic that loses a trained provider pays for recruiting and training, and also for the clients who follow that provider out the door.
Part 3 | Building the weekly schedule
A demand-based schedule has two layers. The core is the set of shifts each person works every week, built to cover the quietest version of each day; the flex layer is a smaller set of shifts added for the peaks, offered first to people who want more hours, and posted well before the week begins.
| Stylists on payroll (Assumption) | 10 (5 full-time, 5 part-time) |
|---|---|
| Chairs (Assumption) | 10 |
| Eight-hour shifts per week (Tue to Sat) (Assumption (5 × 5 + 5 × 2)) | 35 (7 a day) |
| Scheduled staff hours per week (Calculated (35 × 8)) | 280 |
| Booked service hours per week (Assumption (34 + 36 + 46 + 54 + 56)) | 226 |
| Saturday hours turned away per week (Assumption (front-desk tally)) | 10 |
| Total demand per week (Calculated (226 + 10)) | 236 hours |
| Staff hours needed at 85% (Calculated (236 ÷ 85%)) | 277.6 |
| Proposed shifts (Tue to Sat) (Assumption, totals 35 shifts or 280 hours) | 5, 5, 7, 8, 10 |
| Average revenue per service hour (Assumption) | $90 |
| Saturday demand turned away per year (Calculated (10 × $90 × 50 weeks)) | $45,000 |
Illustrative example; replace every assumption with your own numbers.
The proposed schedule moves 32 hours, 16 each from Tuesday and Wednesday, to Friday and Saturday, which brings Saturday to 80 hours against 66 hours of demand, about 83% full. In practice that means four part-time shifts move from midweek to the weekend, so part-time stylists work mostly Fridays and Saturdays while full-time stylists keep their five days, a trade best worked out with the people involved rather than announced.
The routine to keep it working fits on one page:
- Post the schedule two weeks ahead, the practice the retail pilot made standard, so people can plan childcare, classes and second jobs.
- Keep core shifts the same week to week, changing them only at a set time such as each quarter.
- Offer flex shifts first to people who have asked for more hours.
- Let staff swap shifts with each other inside clear rules (same skills, owner sees the swap), which removes the owner as the bottleneck.
- Review the demand curve monthly, comparing booked and turned-away hours with scheduled hours by day.
Predictive scheduling laws apply to some employers. Seattle's Secure Scheduling Ordinance covers hourly employees at retail and food service businesses with 500 or more employees worldwide and requires schedules posted at least 14 days in advance, with extra pay for many later changes, and Oregon's law covers large retail, hospitality and food service employers with at least 500 employees worldwide, also with 14 calendar days' notice. Most owner-run businesses fall below those thresholds, but rules differ by city and state, so check with your local labor agency. General information, not legal advice; talk to a licensed professional about your situation.
How you compare
- Hours matched to demand: in the retail pilot, only 30% of weekly payroll swings followed customer traffic; compare your own scheduled hours with your booked and turned-away hours by day to see how closely they track.
- Notice: two weeks of notice was standard practice in the pilot company and is the legal minimum for covered employers in Seattle and Oregon.
- Your own baseline: the gap between scheduled and needed hours for each day over the last four weeks, and the number of shift changes made after the schedule was posted.
Questions to ask this week
Which day has the most idle staff hours? Compare scheduled hours with booked hours for each day.
How many requests did we turn away, and when? Keep a tally at the front desk for one or two weeks.
Who would trade a weekday for a weekend day? Ask before rewriting anyone's core shifts.
How far ahead do we post the schedule? If the answer is days, move it to two weeks.
The next step is to put four weeks of booked hours and a week of turned-away requests into a day-by-day table and set it beside the schedule you are running now.
About this research: This article draws on the Stable Scheduling Study, a randomized pilot in 28 retail stores, a study of schedule instability and turnover in the ILR Review, a queueing analysis of service capacity, and official pages from the City of Seattle and the Oregon Bureau of Labor and Industries. The salon example is illustrative.
Sources
- Williams, J. C., Lambert, S. J., Kesavan, S. et al. Stable Scheduling Study Report. Center for WorkLife Law, University of Chicago, UNC Kenan-Flagler, 2018. 28 stores, randomized.
- Choper, J., Schneider, D. and Harknett, K. Uncertain Time: Precarious Schedules and Job Turnover in the US Service Sector. ILR Review 75(5):1099–1132, 2022. 1,827 workers.
- Green, L. Queueing Analysis in Healthcare. In Hall, R. (ed.), Patient Flow, Springer, 2006.
- City of Seattle Office of Labor Standards. Secure Scheduling Ordinance.
- Oregon Bureau of Labor and Industries. Predictive Scheduling.
Related reading: How to Run a Business With High Staff Turnover: Training, Checklists and Systems · Turnover Rates by Industry, and What Each Departure Costs · Holiday Scheduling for Salons: Hours, Staffing and Saying No · Capacity Planning for Service Businesses: How Full Is Full? · Utilization Rate: The Number Behind a Busy Week That Still Loses Money
Work with osNOVA on staff scheduling
We help owner-run service businesses turn booking data into a weekly schedule that puts people where customers are, with stable core shifts the team can plan around and a simple flex layer for the busy days, starting from what you want more of, whether that is fuller weeks, fewer turned-away customers or a team that stays.
Email hello@osnovaops.com about your business
Or leave your email and we will reach out
Tell us what kind of business you run and what you want more of.
Not ready to talk? osNOVA runs free live events for owners.