What Is Employee Scheduling? A Guide for Businesses With Delivery Routes

Employee scheduling is the work of deciding how many people you need, which people they are, and which hours they work. Every business with hourly staff does it. The difference, when your day ends with loaded vans going out, is that some of those hours aren’t yours to move.

A 10 a.m. drop window is fixed by the customer. The hour of loading before it is fixed by physics. Build a schedule that ignores either one and you don’t get a mildly inefficient week, you get late deliveries and a driver on overtime. That’s what makes scheduling in a delivery-running business its own problem rather than a generic admin task.

This guide covers the whole practice: what it includes, the six steps to produce a schedule, how to keep labor cost honest, how much notice you owe people, and the point at which a spreadsheet stops being enough. Each shift pattern and tool gets a short treatment here with a link to the full version, starting with how to choose shift scheduling software once you outgrow doing it by hand.

The Bottom Line

• Employee scheduling answers three questions in order: how many people, which people, and which hours. Skip the first and you’re negotiating with your own staff instead of planning.
• Delivery windows are anchor points, not preferences. Schedule backward from the drop time through loading and prep, and the rest of the day arranges itself around those fixed pegs.
• Most small operators are not covered by predictive scheduling laws. Oregon’s rule, for example, applies to retail, hospitality and food service employers with 500 or more employees worldwide (Oregon BOLI, retrieved 2026-10-02). The notice still matters commercially, even when it isn’t legally required.
• Non-daytime work is normal in this part of the economy. 26.0% of workers in transportation and utilities and 25.4% in wholesale and retail trade usually worked a non-daytime schedule, against 16.4% of all workers (U.S. Bureau of Labor Statistics, 2017-18).
• A spreadsheet is fine up to roughly 10 people on stable hours. What breaks it is change: swaps, call-outs and time-off requests arriving faster than you can re-publish the file.

What employee scheduling actually covers

Employee scheduling is the process of matching people to working hours so that the work gets covered without paying for hours you don’t need. In practice it bundles four jobs that owners often treat as separate: forecasting demand, collecting availability, assigning shifts, and handling the changes that arrive after you’ve published.

The assignment itself is the smallest of the four. Most of the pain lives in the forecast at the front and the churn at the back.

Scheduling also gets confused with three systems it is not. It isn’t tracking employee hours, which records what actually happened. It isn’t payroll, which pays for it. And it isn’t route planning, which sequences stops rather than people. Those three systems consume the schedule, so they’re often sold alongside it, but a schedule is a plan about people and hours and nothing else.

Why delivery routes change the scheduling math

In most businesses, hours are elastic and demand is a curve. In a business with delivery routes, part of the day is nailed down. A wholesale customer’s receiving bay closes at 11 a.m., a florist’s Valentine’s orders have to land before the workday starts, a caterer’s drop is tied to an event start time.

That changes the order of operations. Instead of starting with “who’s available Tuesday”, you start with the drops, then work backward:

• Drop window sets the arrival time.
• Route time between the first and last stop sets the departure time.
• Loading and checking sets when the van has to be packed.
• Prep and production sets when the kitchen or bench starts.
• Everything else covers counter duty, admin and cleaning, and fills in around those four.

Work in that order and the early-morning shift stops looking arbitrary. It’s the only shift that can produce a 10 a.m. delivery. Try to schedule forward from opening time instead and you’ll discover the conflict at 9:40 a.m., which is far too late to fix it.

One more wrinkle: driver hours and production hours scale differently. Double your orders and you may need one more baker but three more delivery hours, because distance doesn’t shrink. Forecast the two separately, even if the same people do both.

The six steps to make a work schedule

Here’s the sequence that produces a schedule you can publish. The steps are the same whether you do them in a notebook or a tool, which is the point of learning them in this order.

1. Forecast the coverage you need

Start from volume, not from headcount. For each day, write down the orders or covers you expect, the number of routes going out, and the hours each takes. Past sales data is ideal. Rough patterns are usable: “Saturday is double Tuesday” and “the 7 a.m. route always needs two people loading” both count as inputs.

Convert that into required hours per role per day. This is the number everything later gets measured against.

2. Collect availability and time-off requests in one place

Availability gathered in four channels (text, a shouted conversation, a note on the board, a message to a supervisor) is the single most common cause of a broken schedule. Pick one place and insist on it. Set a weekly cutoff for requests so that late asks are the exception you choose to accommodate rather than the default.

Record skills and licences alongside availability. Three people free on Thursday is useless if none of them can drive the van.

3. Assign the constrained shifts first

Fill the hardest slots before the easy ones: the early route, the one weekend shift nobody wants, the hours only two people are certified for. Those are where your schedule can actually fail. The flexible mid-week daytime hours can absorb whatever’s left.

If one of those hard slots keeps landing on the same person, you have a fairness problem that will turn into a turnover problem. Rotating the unpopular hours is the standard answer, and rotating shifts spread the load but cost sleep is the trade to weigh before you commit to a pattern.

4. Price the schedule before you publish it

Multiply the hours by the loaded rate, which is the wage plus payroll taxes and benefits rather than the wage alone, and compare it against expected revenue for the week. Do this while the schedule is still editable. A schedule you can’t afford is much cheaper to fix on Thursday than to discover on payday.

Watch for anyone crossing into overtime, and watch for the quiet version: a person scheduled 38 hours whose shift reliably runs 20 minutes long five times a week.

5. Publish the schedule two weeks ahead

Two weeks is the standard to aim at, for reasons covered further down. Publish somewhere everyone can see the current version, and make it obvious which version is current. A schedule that lives in three WhatsApp messages isn’t published.

Reminders are the cheap companion to notice. A nudge the day before a shift costs nothing and prevents the honest kind of no-show, which is the same reason appointment-based businesses run appointment reminder software against their own bookings.

6. Build a path for changes

Changes aren’t a failure of the schedule, they’re a feature of employing people. Decide in advance who can approve a swap, how a call-out gets covered, and where the result gets recorded. Teams that let staff swap among themselves inside clear rules spend far less manager time on this, which is the practical form of flexibility for hourly teams.

How to keep labor cost in line without short-staffing the route

Labor cost percentage is total labor divided by sales for the same period. It’s the fastest check on whether a schedule is defensible, and it’s worth tracking weekly rather than monthly, because a monthly number tells you about a problem you can no longer fix.

What counts as a good percentage depends entirely on the model. In food service, full-service restaurants have been running a median around 36.5% of sales while limited-service and quick-service models sit closer to 25-31% (Restaurant Inventory Tools, 2026). A wholesale baker with three delivery routes has a different structure again, because a chunk of labor is production rather than service.

Two traps are specific to delivery:

• Cutting the loading shift. It’s the easiest hour to delete and the most expensive to lose. An unloaded van at departure time costs the route, not the hour.
• Treating overtime as the flex. Overtime is the most expensive hour you can buy, and it tends to land on your most capable person, which is also the person most likely to leave.

The structural fix is matching shift lengths to the actual shape of the day. If the work is a morning production block and an afternoon delivery block with a dead middle, two shorter shifts beat one long one. If you’re in a restaurant, the same logic runs against service rushes rather than routes, which is covered in the guide to restaurant staff scheduling.

How much notice do you owe employees?

Legally, in most of the United States, none. Predictive scheduling laws exist in a handful of jurisdictions and most of them are aimed at large employers. Oregon’s statewide rule covers retail, hospitality and food service employers with 500 or more employees worldwide, requires the written schedule at least 14 calendar days before the first day on it, and requires a rest period between shifts (Oregon BOLI, retrieved 2026-10-02). Seattle, New York City, Chicago, Philadelphia, San Francisco and a few others have their own versions with their own thresholds, and New York City’s retail rule runs on 72 hours’ notice rather than 14 days.

So if you have 12 employees, this probably isn’t a compliance question for you. Treat it as a retention one instead. The reason the laws exist is that short-notice scheduling costs workers money and sleep, and the workers most able to leave over it are the ones with licences and skills you can’t quickly replace.

A workable standard for a small operation: publish two weeks out, treat the first of those weeks as firm, and pay attention to how often you break your own rule. Three changes a week isn’t a notice problem, it’s a forecasting problem.

When a scheduling spreadsheet stops being enough

A spreadsheet handles scheduling well at small scale, and there’s no virtue in buying software before you need it. The honest signals that you’ve outgrown it:

• You’re re-sending the file because someone is working from last Tuesday’s version.
• Swaps and call-outs are arriving faster than you can update it.
• You don’t know your labor cost until payroll tells you.
• Time-off requests live in a different place than the schedule.
• Someone hit overtime and nobody noticed until it was paid.

Here’s roughly where each approach gives out:

The last row matters. Shift tools and route tools solve adjacent problems and neither substitutes for the other, which is why operators at scale usually run both. The second half of that pair is delivery scheduling software, which handles drops and windows rather than people and hours. Entry pricing on the shift side is low enough that cost rarely decides it: When I Work lists Essentials at $2.50 per user per month, Deputy’s Lite plan is $5 per user per month with a $30 monthly minimum, and Homebase’s Basic plan is free for one location and up to 10 employees (When I Work, Deputy, Homebase, retrieved 2026-10-02).

Common employee scheduling mistakes

• Scheduling from headcount instead of demand. “Four people on Saturday because we always have four” is a habit, not a forecast.
• Publishing a schedule nobody priced. The cost is knowable before the week starts.
• Letting the schedule and the time-off list live apart. This is how you schedule someone who told you in March they’d be away.
• Assuming the van packs itself. Loading is work and needs hours on the schedule.
• Giving one person all the awkward shifts. It’s cheap this month and expensive when they quit. Recruiting, onboarding and training a replacement driver or baker costs far more than the few hours it takes to rotate the hard slots fairly.

Frequently asked questions

What is employee scheduling in simple terms?

It’s deciding who works which hours, and making sure the hours you need covered are covered without paying for hours you don’t. The full practice includes forecasting how many people you need, collecting availability, assigning shifts, publishing the result, and managing changes afterward.

How far in advance should a schedule be published?

Two weeks is a good target and is the standard several predictive scheduling laws codify for large employers. Most small businesses aren’t legally required to meet it, but longer notice reduces call-outs and turnover, both of which cost more than the planning time does.

What’s the difference between employee scheduling and workforce management?

Scheduling is one function: people to hours. Workforce management is the broader bundle that usually includes scheduling plus time tracking, attendance, leave, labor forecasting and sometimes payroll. Most scheduling apps have grown into light workforce management tools.

How do you schedule staff when demand is unpredictable?

Schedule a reliable core to your minimum expected volume, then keep a documented way to add hours: a short on-call list, open shifts staff can claim, or a standing agreement with two people who want extra hours. The goal is a planned response to a surge rather than a scramble.

Can one schedule cover both drivers and production staff?

Yes, and it usually should, because the same people often do both jobs. Forecast the two kinds of hours separately, because production scales with order count while driving scales with distance and stops, then combine them into one published schedule so nobody is accidentally booked in two places.