How to create a demand based schedule

When you build rotas from last week’s schedule instead of this week’s demand, you risk overstaffing when it’s quiet and scrambling when it’s busy (with the labour bill to match).

As we all know, there’s not much room for error when it comes to labour costs. 

Demand-based scheduling fixes that by matching your staffing to when customers are actually likely to arrive. But how does it work? And how can you incorporate demand-based scheduling into your restaurant? 

This guide explains all these answers. By the end, you’ll know what a demand-matched schedule is, how to build one, and what you need to know before you get started.

Let’s start with the basics. 

So, what is demand-based scheduling in a restaurant? 

A demand-based schedule means building the staff rota from forecasted demand rather than estimates or gut feel. The forecast helps you figure out how many people you need, by role, hour by hour, so you can build an optimal rota. 

The process replaces pattern scheduling, where the same rota repeats weekly with manual adjustments when someone notices it's wrong. 

How does demand-based scheduling work? 

Demand-based scheduling means putting the right number of people in the right roles at the right time. To do that, you forecast demand, turn it into staffing requirements, build the rota, then adjust it as actual trade comes in.

Here’s what that looks like in practice: 

1. Forecast demand by day part. A forecast that only tells you Saturday will be busy isn’t much use. You need to know when it’ll be busy throughout the day so you can schedule staff accordingly. Start by estimating covers or transactions in blocks you can schedule against, like hourly. Use your sales history, then account for things that can change demand (weather, local events, holidays, and promotions are some examples). 

2. Turn demand into a labour requirement. Next, translate that forecast into people. If you expect 90 covers between 7pm and 8pm, how many chefs, servers, and bar staff do you need? This is where blanket labour ratios (also known as one-size-fits-all ratio) can be tricky to use. The kitchen might need three people for that level of trade, while the floor needs five. You can’t always use the same rule for both.

3. Build the rota around your people. Now match those requirements to your actual team. That means considering availability, skills, contracted hours, shift lengths, and who can run each section. Nory’s Scheduling Assistant can do this automatically, turning the labour plan into a rota that works with the people you actually have.

4. Check compliance as you build. The rota also needs to work within the rules. That can include break entitlements, rest between shifts, young worker restrictions and, in some markets, scheduling notice requirements. Nory’s AI Compliance Assistant can help here, automatically checking regulations when building rotas.

5. Publish, then adjust against actual trade. Once the week starts, compare planned hours with what’s actually happening and adjust shifts where needed. 

Recommended reading: Why AI forecasting is essential for restaurant success. 

Demand-based scheduling vs fixed scheduling: How do they compare? 

The main difference is what you build the rota around. A fixed rota follows a familiar pattern, while a demand-based rota changes with the forecast. One prioritises predictability, the other focuses on having the right number of people for the trade you expect.

Here’s how they compare: 

Fixed rotas often seem more appealing because they’re predictable. A rota that changes shape weekly is unpopular in a sector that already struggles to retain people, so having some stability for your core team can be a benefit for staff morale and retention. 

But the reality is that fixed rotas aren’t sustainable when demand fluctuates, which it usually does. If you stick with a rigid schedule when demand changes, you risk overspending on labour costs and cutting into your margins. 

How much does demand based scheduling reduce labour cost?

For Nory customers, demand-driven scheduling typically cuts labour costs by 10 to 20% in the first eight weeks. However, the real saving depends on how far your current rota is from actual demand. 

If you’re already scheduling tightly against covers, there’s less to gain. If you’re copying last week’s rota forward, there’s usually a much bigger gap to close.

There are some useful real-world examples. Josie’s cut labour costs by 23% in four months when using Nory to create demand-based schedules, while Passyunk Avenue reduced labour costs by 26% across its sites. 

Digbeth Dining Club, which runs a more complex multi-venue and events operation, also keeps labour within 0.38% of plan using Nory. 

Now we know exactly when people are in the building spending money. That changes everything.
Nicol Dwyer, Operations Director at Digbeth Dining Club

Demand-based scheduling doesn’t magically create savings, but it does prevent you from paying for hours you don’t need and avoid the expensive scramble when demand is higher than expected.

And because labour is one half of prime cost, it’s often the first place operators look when they want to improve margins.

When demand-based schedules can fall short

Demand-based scheduling works best when the forecast is good, the rota is flexible, and managers can still use their judgement. 

There are four common things to watch for: 

• The forecast doesn’t add enough value. An experienced general manager who’s run the same site for years will know things a system might miss. The point isn’t to replace that judgement, but to give it better data to work with. Compare planned labour with actual labour each week. If the gap isn’t getting smaller, look at the forecast and the assumptions behind it.

• The rota is set too far in advance. Staff need notice, but demand forecasts get more accurate as the day gets closer. Publishing everything three weeks ahead can lock you into a plan that no longer reflects what’s expected. A better approach is to set core shifts early, then confirm flexible shifts closer to the day.

• The rota doesn’t account for real-world constraints and compliance. Breaks, rest periods, contracted hours, and skills all affect who you can actually schedule. If you only check these after building the rota, you’ll end up making manual changes anyway. Build those constraints in from the start and the plan is much more likely to work (which you can do with Nory’s Compliance Assistant).

• Managers keep overriding the schedule. An override isn’t necessarily a problem. Like we said, managers might know about things a forecast won’t, like a local event. But if the same changes happen every week, that’s useful information that could mean staffing rules need adjusting. Review those overrides regularly and use them to improve the next schedule.

Recommended reading: The future of restaurant compliance: From manual checks to AI assistants.

What you need before building a demand-based schedule

Before you can build a demand-based schedule, you need four things: useful sales data, a forecast you can measure, clear information about your team, and the rules they work under. Without those basics, you’re still relying on guesswork, just with a few more steps.

Let’s look at these in more detail: 

1. Sales history you can schedule against. You need sales data broken down by day part (ideally into hourly or half-hourly blocks) to give something useful to forecast from. If all you know is that Saturday is busy, you can’t work out when you actually need more people.

2. A forecast you can measure. You need to know how accurate your forecasts are and whether they’re getting better over time. Otherwise, there’s no way to tell if your new rota is actually an improvement. accurate, consistent, and detailed enough to forecast demand by day part (which is where a system like Nory can help). 

3. Clear information about your team. Roles, skills, availability, and contracted hours all need to live somewhere reliable, like a restaurant operating system. If it doesn’t, even a good labour plan can produce a rota nobody can actually work. 

4. Your compliance rules built in. Breaks, rest periods, contracted hours, and any local scheduling requirements need to be part of the process from the start. If you check them after building the rota, you might end up rebuilding it by hand. Or worse, miss a key requirement and publish a schedule that isn’t legally compliant.

The first two are the foundation. Without the data and a measurable forecast, there’s nothing meaningful to schedule against. The last two make the plan workable in the real world.

How Nory helps operators with demand-based scheduling

Nory is an agentic AI restaurant operating system, with a crew of AI Assistants that manage operational areas against a prime cost target. The software works best for multi-site operators who want to connect demand forecasting and scheduling, so the rota can respond when demand changes. 

But how exactly does it work? 

The Forecasting Assistant predicts demand at around 97% accuracy and feeds that forecast directly into the Scheduling Assistant. When the forecast changes, the rota can change with it, without someone having to re-enter the numbers or rebuild everything from scratch. 

The Compliance Assistant also applies relevant rules when building the rota rather than leaving managers to check them afterwards.

The results? Better schedules that meet demand and lower labour costs. 

Take a look at Barge East as an example. After using Nory’s forecasting, the multi-site restaurant cut its labour costs by 10%: 

Recommended reading: The reinvention of workforce planning with restaurant labour scheduling AI.

FAQs about demand-based scheduling

How to build a schedule based on forecasted demand?

Start by forecasting demand in manageable blocks, usually hourly or half-hourly. Then turn that forecast into a staffing requirement for each role, match it to your team’s availability, skills and contracted hours, and check the relevant compliance rules. Once the rota is live, compare planned labour with actual trade and adjust the remaining shifts.

The important part is connecting your demand forecast to the rota. You’re not just predicting how busy you’ll be, you’re using that prediction to decide how many people you need, when you need them, and where.

What data do you need for demand-based scheduling?

You need:

• Historical sales data
• A forecast of expected demand
• The number of staff needed for each role
• Staff availability
• Contracted hours
• Relevant employment and scheduling rules
• A way to compare planned staffing with actual demand

How accurate does a sales forecast need to be for scheduling?

There’s no magic accuracy percentage, but the more accurate your data (and the more often it updates), the better your labour planning will be. 

Does demand-based scheduling work for small restaurants?

Sometimes, but it isn’t always worth the effort. If you run one site with stable trade and an experienced manager who knows the business inside out, a simple spreadsheet may be enough. 

However, relying on one person to create schedules can be a risk. If they’re away, leave, or simply make a bad call, that knowledge goes with them. 

The case gets stronger as you grow. Once you have multiple sites, you need a consistent way to forecast demand and build rotas without relying on what one manager happens to know about their site.

Use Nory for demand-based scheduling without the guesswork

Demand-based scheduling works when you have the right data, a useful forecast, and a rota that accounts for your actual team and constraints. The hard part is connecting all three and keeping the schedule responsive as demand changes. 

That’s where Nory can help, bringing forecasting, scheduling, and compliance into one system.

Get in touch with the team to start building demand-matched rotas in under 5 seconds. 

Disclosure and methodology

This page is published by Nory, an agentic AI restaurant operating system with demand-matched scheduling. That means we have a commercial interest in the approach. We’ve tried to keep the analysis honest, including where demand-based scheduling can fall short.

Customer figures come from published Nory success stories. Where we describe typical outcomes, we say “customers typically see” because results depend on how closely your current rota matches demand.