Demand and sales forecasting

Restaurant demand forecasting, with the reasoning shown

A forecast you can't question is a forecast nobody uses. Orbis predicts how many guests you will have and what they will order, per dish and per service, and shows exactly why: last four Fridays, rain from 14:00, a concert 1.2 km away, 38 tables already booked. Then it turns the number into prep, orders and staffing.

What Orbis reads before it forecasts

  1. 01

    Your POS history

    What sold, when and how fast, on every channel, backfilled from the POS you already use.

  2. 02

    Weather

    Rain, sun, cold and snow move covers and the menu mix. A cold day lifts soup; a terrace day moves everything outside.

  3. 03

    Bookings

    Tables already booked against what is usual for that day.

  4. 04

    Local events

    Concerts, matches and festivals nearby, and how far away they are.

  5. 05

    Calendar

    Day of week, time of month and school term or holidays.

Demand forecasting and sales forecasting: the difference

Restaurant sales forecasting predicts revenue. Demand forecasting predicts what drives it: covers, and how many of each dish. The second is what a kitchen can act on, because prep, ordering and rotas are made of portions and hours, not euros.

Orbis forecasts demand per dish and rolls it up to covers and revenue, so the same number serves the chef, the buyer and the owner.

From forecast to action

The forecast is the start of a loop that runs before the kitchen opens:

  • Prep: a misa list of quantities per station, netted off what is already made
  • Inventory: confirmed prep checked against live stock, shortfalls flagged
  • Ordering: a draft cart per supplier at your invoice prices
  • Labour: staffing need per hour and labour cost before the shift starts

Shadow mode: proof before trust

Orbis earns trust before it asks for it. In shadow mode it runs the whole loop silently beside your kitchen, predicting every service and checking itself against what actually happened. You see its accuracy week by week and decide when to let it suggest, then assist.

Every override you make is logged as training data, so the forecast learns your kitchen rather than an average one.

Questions

What is restaurant demand forecasting?

+

Predicting how many guests a restaurant will serve and what they will order, per service, so prep, purchasing and staffing can be planned in advance instead of guessed.

How accurate is the forecast?

+

It depends on your kitchen, which is why Orbis starts in shadow mode: it predicts every service and compares itself with what actually happened, so you can see its accuracy before you rely on it.

Is it a black box?

+

No. Every forecast comes with the reasoning: which signals moved it and by how much. Overrides are welcome and the model learns from them.

What if my restaurant is new?

+

Orbis works from your own POS history today. A day-one forecast for new restaurants, built from anonymised patterns across kitchens, is on the later roadmap.

Which POS does it work with?

+

Not on day one. Orbis starts by reading the sales history of the POS you already use, read-only, and runs in shadow mode beside it. Its own till, CASA, which runs on any tablet and keeps working offline, is planned for January 2027, so you can switch when you are ready. When you apply, tell us what you run today: Restolution, Trivec / Caspeco, Lightspeed, Toast, Square or another.

See it run on your kitchen.

R&D access is free during testing. Orbis runs in shadow mode first, with no risk to service. We read every application and reply within 48 hours.

Apply for R&D access →