ContactBook a call

Industries/ Hospitality/ Rota and demand forecasting

Demand forecasts that show the covers coming, beside your rota

We forecast covers per site, day and meal period from your own trading history, your bookings, the weather and what is happening nearby. The forecast sits beside the rota as a range, early enough for a manager to act on it, and the manager still decides.

A restaurant manager, a man in his thirties with tattooed forearms and a black chef's jacket, at the pass before service reviewing next week's rota on a tablet, chefs preparing food behind.
Used by
General managers, area managers and operations directors
Takes the place of
Last year's diary, gut feel and wage surprises after payroll
Connects to
Your till, reservations, rota software and a weather feed
Built with
Our forecasting & alerts service

Where it hurts today

A normal week, before it exists

On Tuesday morning the general manager writes the rota for the week after next. Open on the desk are last year's diary, the bookings so far and a memory of how the same week went. The football fixture, the half-term week and the warm spell nobody saw coming are in someone's head or nowhere at all, and the rota has to reach the team by Wednesday so people can plan their lives.

The cost of getting it wrong shows up late. A quiet Sunday with eight on the floor eats the margin from the whole weekend, while a Friday two servers short means slow tables, a stretched kitchen, a tired team and a review about waiting forty minutes for mains. Neither is obvious until the week is over and the wage report arrives, by which point the money is spent and the shifts are history.

For a hotel, the same problem appears at breakfast and in housekeeping. Breakfast covers follow occupancy and the mix of business and leisure guests, and housekeeping hours follow departures and stayovers rather than arrivals. The property management system knows all of it, but it is not where the rota is written, so the head housekeeper works it out from a printout every morning.

What it does

Rota and demand forecasting, feature by feature

What the forecast does, and how it fits into the way your managers already write rotas.

  1. 01

    Covers forecast per site, day and meal period

    Breakfast, lunch and dinner are forecast separately for each site, as far ahead as your rotas are written, from your till history and the bookings already on the book. As more bookings arrive, the forecast for the coming days firms up and the range narrows.

  2. 02

    A forecast range, not a single number

    Each shift shows a likely low, a likely high and the most likely figure. A tight range means the forecast is fairly sure. A wide one means plan a flexible shift, someone on call or a later start, rather than betting the wage bill on one number.

  3. 03

    Labour forecast beside the rota, by role

    The forecast sits next to the hours already planned for floor, bar and kitchen, so a manager sees at a glance where the rota looks short or heavy against likely demand, while there is still time to change it before it goes out.

  4. 04

    Local events, weather and holidays

    Bank holidays, school holidays and your own events are built in. Managers add local fixtures, concerts and conferences to a shared calendar, and the weather forecast adjusts the coming days for sites where a terrace or a beer garden changes everything.

  5. 05

    Breakfast and housekeeping forecasts for hotels

    For hotels, breakfast covers come from expected occupancy and guest mix, and housekeeping hours from departures and stayovers in the property management system, so the head housekeeper and the breakfast lead plan from the same picture as reception.

  6. 06

    Overrides with a reason, and honest accuracy

    Managers can override any shift, for example for a private party the forecast cannot know about, and add a short note. Each week the forecast, the override and the actual covers are compared side by side, so everyone can see which was closer.

A day with it

How the day runs instead

A rota week at a four-site restaurant group with the forecast running.

  1. Tuesday, 9.30am
    General manager

    Opens the week after next. Friday dinner shows a likely 140 to 175 covers against a rota built for 120, so two more servers and a kitchen porter go on. Sunday dinner looks heavy, so one floor shift becomes a later start.

  2. Tuesday, 9.45am
    General manager

    Adds the wedding party booked into the private room on Saturday, which the forecast cannot see yet, as an override with a note. Publishes the rota to the team from the usual rota software, a day earlier than last month.

  3. Wednesday
    Area manager

    Compares the four sites' rotas against their forecasts on one screen. One site has planned heavily for a quiet Monday. A short call with that manager explains it: a training session, which goes into the notes.

  4. Saturday, 2pm
    Duty manager

    Rain is now forecast all evening, and the dinner range at the terrace site drops. The duty manager offers an early finish to the two people who had asked for one, rather than deciding at 8pm with the floor half empty.

  5. Monday
    Operations director

    Reviews last week: forecast, override and actual covers for every shift, and planned against worked hours for each site. Friday's forecast was close. Saturday's override was closer than the forecast. Both are noted for the next run.

What it connects to

Works with what you already run

The forecast learns from the systems you already run, and shows its result where your managers already plan.

  • Till and point of sale

    Covers, sales and the time of each order, per site and meal period, going back as far as the till keeps them. This is the history the forecast learns your weekly and seasonal patterns from.

  • Reservations or property management system

    Bookings already on the book for coming dates, large parties and, for hotels, occupancy, arrivals and departures. For the next few days, these are often the strongest signal of all.

  • Rota and time and attendance software

    Planned shifts by role and the hours actually worked, so the forecast can sit next to the rota, for example Planday or Deputy, where they offer an API. Otherwise the forecast is shown on its own screen beside it.

  • Weather and local events

    A weather forecast service for each site's location, plus the shared event calendar your managers keep: fixtures, concerts, conferences and markets that move demand near a particular site.

The first version

Start small, then grow it

A first version should be useful on day one and small enough to get right. This is where we would usually draw the line, and what tends to follow once people are using it.

In the first version

  • Covers forecast per site, day and meal period, shown as a range
  • Forecasts tested against your own history before anyone relies on them
  • A side-by-side view of forecast covers and planned hours by role
  • A shared calendar for local events and private bookings
  • A weekly comparison of forecast, override and actual covers

Once it is in use

  • Hotel breakfast and housekeeping forecasts from the property management system
  • Weather adjustments for sites with terraces or outdoor seating
  • Suggested hours per role from the forecast, for managers to accept or change
  • Alerts when a coming shift moves well outside the range planned for it

Buyer's guide

Before you commission one

Four questions worth answering before you commission a forecast.

  1. 01

    Your rota software's forecast or a separate one

    Some rota and workforce management products include a basic sales or labour forecast. If yours does and your managers trust it, improving what feeds it may be all you need, and we will say so. A separate forecast earns its place when you need covers by meal period, bookings, events and weather combined, hotel occupancy, or one honest view across several sites.

  2. 02

    Gather the history and the odd weeks

    For yearly patterns such as summer, Christmas and the January dip, two full years of till data is a comfortable minimum. Gather the till history by meal period, the booking history if you have it, and a list of the odd weeks: refurbishments, closures and one-off festivals. A new site can borrow patterns from a similar site until it has its own.

  3. 03

    Agree who has the final word

    General managers own their rotas, so they keep the right to override the forecast, and overrides are compared with the result rather than criticised. Someone central, usually the operations director, owns the forecast itself: reviewing accuracy each month and asking for changes. A forecast managers feel judged by gets ignored. One they can argue with gets used.

  4. 04

    Measure it against how you do it now

    Over a few months, compare the forecast with the way rotas are written today, usually the same week last year plus a judgement. Watch forecast against actual covers per shift, how often managers override and who was closer, labour against sales by site, and last-minute shift changes. If it is not closer than your current method, we will say so.

Terms worth knowing

The words you will meet

Covers
The number of guests served in a meal period, usually counted from the till or from bookings. A forecast of covers is more useful for staffing than a forecast of sales, because it is people, not takings, that need serving.
Backtesting
Running the forecast over past weeks where the real answer is already known, to see how far off it would have been. It is how you find out whether a forecast deserves trust before anyone writes a rota from it.
Predictive scheduling
Laws in some places, such as Oregon, that require certain employers to publish schedules in advance and generally pay extra for late changes. Rules differ from place to place. They matter because a forecast only helps if it arrives before the rota deadline your rules set.

Handled with care

The details that matter here

A forecast affects people's shifts and wages. These are the parts we handle with particular care.

  • Managers decide, the forecast advises

    We build the forecast to inform the rota, not to write or cut it on its own. Changes to shifts stay with the manager, inside the rota software your team already uses, and the reasons behind each forecast are visible so a manager can explain a decision to the team.

  • Gaps and odd weeks in the history

    Refurbishments, closures, a burst pipe or a one-off festival all distort the history. We go through the odd weeks with you and mark them, so the forecast does not expect a closure every March, and we tell you if the history is too short or patchy to forecast well.

  • Wages and personal data

    The forecast needs hours and roles, not names or pay rates. Where labour cost is shown, we agree with you who sees it, usually each manager for their own site, and individual pay stays in the payroll and rota systems where it already lives.

  • Shift notice rules where you operate

    Notice rules differ by country, state and city. In Oregon, for example, large retail, hospitality and food service employers must give written schedules 14 days ahead. We design around the rules that apply to you, agreed with whoever handles employment matters, and the forecast is there to help managers plan earlier, never to justify last-minute cancellations.

Questions

What buyers ask first

Can demand forecasting write our staff rota?

No. It shows the covers likely for each shift beside the rota your manager is writing, and later it can suggest hours per role for the manager to accept or change. Who works which shift, who has asked for a day off and who is ready to run the pass on a Saturday are decisions that stay with the people who know the team.

How accurate is restaurant demand forecasting?

It depends on your data, and nobody can honestly quote a figure before seeing it. We backtest against your own history and show you how far off the forecast would have been. Steady, well-recorded sites tend to forecast well. A sudden storm or a coach party will still catch it out, which is why it shows a range, managers can override it and misses are reviewed weekly.

How much sales history do we need for forecasting?

Two full years for summer and Christmas patterns, though less still helps. With a year of data, weekly patterns and the bookings already on the book give a useful view of the coming days, though the seasonal picture is less certain. A new site can borrow patterns from a similar site in the group until it has history of its own. We show you how well it does before you rely on it.

Does our rota software already include labour forecasting?

Possibly, and if it is good enough and your managers use it, we would rather help you get more out of it. A separate forecast earns its place when you need covers by meal period, hotel occupancy, events and weather combined, or a comparison across sites that the rota product does not offer.

How will our managers come to trust a forecast?

By seeing it tested, not by being told. The drivers behind each forecast are visible, so a manager can see that Friday is high because of bookings and a concert nearby. Each week the forecast is shown next to the actual covers and the manager's own override. Trust comes from a few weeks of that, and from managers keeping the final say.

Tell us about the shift that caught you out

Tell us how rotas are written today and which systems hold your covers and hours. We reply within 48 hours with an honest view of whether your history can support a useful forecast.