Forecasts next week's usage from sales history, weather, reservations, local events, and seasonality. Tells you exactly how much of what to order and when, so you stop 86ing best-sellers on a Saturday night and stop throwing away overprep on a Tuesday.
Every capability listed below is included in the base module. No add-ons, no unlock fees, no per-seat surcharges. If you're on xRESTAURANTx, you have this.
Weekly, biweekly, and monthly views on the same forecast. Weekly is what you order from. Biweekly and monthly are what you show your GM and your accountant.
Reads the local 7-day forecast. Rainy nights kill patio covers by 40 to 60 percent and spike delivery orders by 15 to 30 percent. Your order quantities move accordingly.
Booked covers for the week pull directly from the reservations module. A private buyout Saturday? The forecast reads it and adjusts prep, protein, and beverage orders.
Concerts, home games, conventions, festivals. If a 40,000-seat event lets out four blocks away at 10 PM, the model knows it and pre-loads your late-night order.
Year-over-year signal is separated from short-term signal. Snowbird season, tourist season, summer break, holidays: all baked into the baseline so you're not surprised by predictable swings.
Every forecast comes with a confidence range. Top sellers get tight intervals. Long-tail items get wider ones so you don't over-trust a shaky number.
The forecast turns into draft POs to the right vendors with one click. Adjust quantities, send. The system remembers your adjustments and gets closer next week.
If forecast plus current on-hand shows you'll run out before the next delivery, the alert fires with days-of-cover remaining. Bump the order, move the delivery up, or trim the menu.
Mirror of shortfall. If demand is coming down and on-hand is high, the alert prompts you to trim the next order, run a special, or transfer stock to another location before spoilage.
Every Monday, last week's forecast is graded against actual usage per SKU. That's how you build trust in the model, and how you spot which items still need human tuning.
Every operator we onboard runs this same tutorial in their first week. Most complete it in one sit-down; the model gets useful inside 30 days of real usage.
Predictive Ordering trains on your last 90 days of sales at minimum, 12 months ideally. If you're new to xRESTAURANTx, we can back-load your prior POS export in one pass.
Toggle on weather, local-event calendar, and reservation lookahead. Each source improves accuracy for a different kind of demand shock.
The first weekly forecast lands on the ordering screen. Sit with it once and note where your gut disagrees. That gut feedback trains the model.
One click turns the forecast into draft POs to the right vendors. Adjust quantities if you want, then send. The system remembers your adjustments and applies them going forward.
Every Monday, the backtest shows last week's forecast vs. actual usage by SKU. That's how you build trust in the model, and how you spot which items still need human tuning.
xRESTAURANTx modules are not sold separately. Bundle everything, you pay one flat rate keyed to your restaurant's size, not a per-module upsell.
Concrete scenarios from actual operators using the module in production. Every scenario carries a metric or a mechanism, not a marketing slogan.
Private event on Saturday: 80 covers, tasting menu, wine pairing. The forecast reads the reservation block, sizes the protein order Wednesday, sizes the wine pull Thursday, and pre-books an extra prep cook Friday. Nothing is left to chef memory or a sticky note.
Restaurant sits three blocks from a high school. Every school day, the 3:15 rush is real. The forecast reads the local school calendar (in-session, half day, break, exam week), sizes the after-school SKUs correctly, and warns when a spring-break week will drop the spike.
A national franchise with 40 units gets a per-location weather-adjusted forecast every Monday. The Miami unit orders more delivery packaging; the Chicago unit orders more patio protein because a warm week is coming. Regional director sees one dashboard.
A virtual brand's signature item goes mildly viral Tuesday night. By Wednesday morning, forecast detects the spike, drafts an emergency PO for the two key ingredients, and warns the operator that current staffing won't cover Thursday. Two decisions get made before lunch instead of at 8 PM in the weeds.
90 days is the practical minimum for a solid weekly forecast. 12 months gives you seasonality. If you have less than 90 days on xRESTAURANTx, we back-load your prior POS export so the model has enough to work with from day one.
For patios, yes. For delivery, absolutely. Rainy nights kill patio covers by 40 to 60 percent and spike delivery orders by 15 to 30 percent. The model reads the local 7-day forecast and adjusts your order quantities so you're not overprepped for a covered patio night.
Add it to the calendar as a private event with expected covers. The model will adjust its forecast around that block. You can also mark a normal service as excluded (say, a slow holiday) so it doesn't drag your baseline down.
For top-30 SKUs, most operators land inside 10 to 15 percent MAPE within the first month. Long-tail items are always noisier, and the model shows a wider confidence interval on those so you don't over-trust a shaky number.
Always. The forecast is a draft, not a mandate. Operators override every week for their own reasons (a new server, a bad review, a competitor closing). The system learns from your overrides, so what you override this month gets right on its own next month.
If the forecast plus current on-hand shows you'll run out of an item before the next scheduled delivery, the alert fires with days-of-cover remaining. You can bump the order or move the delivery up. This is how you stop 86ing best-sellers on Saturday night.
The mirror of shortfall. If the forecast shows a slowdown coming and current on-hand is well above expected usage, you get an alert to trim the next order, run a special, or shift the stock to another location. This is where the model pays for itself in spoilage saved.
xRESTAURANTx modules are designed to work together. These three are the ones most operators use alongside Predictive Ordering.
This page is free to reprint with attribution to https://xrestaurantx.com. Copy the text, paste it into your operator newsletter, franchise handbook, or vendor briefing. Just keep the attribution line and the link back to the source.
Book a 20-minute demo. We'll pull your last 90 days of sales, run a live forecast against next week, and show you what a proper order sheet looks like.