# SYSPRO Forecasting: How Forecasts Feed MRP > Source: https://netray.co/erp-help/syspro-forecasting-mrp-forecast-setup-and-consumption > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: ERP Help ## Short answer In SYSPRO a forecast is a planned demand quantity by stock code and period that MRP reads alongside sales orders. You load or calculate forecast figures, MRP nets actual sales orders against them within a consumption window so demand is not double counted, and the planning run creates suggestions for purchase orders and jobs. ## Kind how-to ## System syspro ## Erp SYSPRO ## Module Forecasting / MRP ## Query syspro forecasting ## Alternate Queries - syspro forecast mrp - syspro forecast consumption sales orders - syspro demand forecast methods ## Applies To SYSPRO 7 and 8 with Material Requirements Planning; forecast options depend on the licensed modules and on release. ## Steps ### How to set up and use forecasting with SYSPRO MRP **Items:** - Decide which stock codes need forecasts: typically make-to-stock and long-lead items, not every part. Make-to-order items often need none. - Establish the forecast: either calculate it from history using a statistical method in your forecasting setup, or import demand from sales or a planner's spreadsheet. - Review forecast quality per item and period. Remove one-off spikes from history and override where sales knows something the history does not. - Set the consumption rules in the MRP setup so actual sales orders reduce the forecast inside the defined window, preventing double counting. - Run MRP and review the output for the forecast items: planned orders, exceptions and suggested dates. - Measure forecast accuracy each month and adjust methods or ownership of the numbers. ## Sections ### How forecast and orders interact **Body:** If MRP simply added forecast to sales orders, demand would be counted twice, because customers eventually order what was forecast. Consumption logic fixes this: as a real order arrives in the forecast window, it uses up a matching amount of forecast, and only the unconsumed remainder plus any extra orders drives supply. The window matters. A short window gives you protection against stale forecast that never gets consumed, while a long one gives smoother planning but carries more risk of inflated demand. Agree the window with sales based on your actual order lead times. ### Choosing a forecasting method **Body:** Statistical methods such as moving averages and exponential smoothing suit stable demand with some noise, while seasonal methods suit products with a calendar pattern. For lumpy, project-driven demand, no statistical method will be accurate; use customer commitments and a planner's judgement. Start simple. A moving average with a clean history beats a sophisticated method fed with dirty history. Segment items by volume and variability, and spend effort where error costs money: high-value, long-lead and constrained items. ### Checking the MRP result **Body:** After a run, read the planning output for a sample of forecast items before releasing anything. Compare gross requirements to what you expect from forecast plus orders and watch for items where consumption did not occur, for example because the order date fell outside the window or the order is for a different warehouse. Keep the forecast owned by someone. A forecast no one reviews becomes a source of excess stock, and the planning exception list becomes noise that planners learn to ignore. ## Pitfalls - Forecasting every item, which creates noise and excess inventory. - Not defining consumption, so MRP double counts forecast and orders. - Using history that includes one-off bulk orders without cleaning it. - Leaving old forecasts in place after a product is phased out. - Not measuring accuracy, so a bad method persists unnoticed. ## Ai Angle Netray adds a private AI layer to the SYSPRO you already run, including on-prem installs where the SQL Server and application server never leave your network. ERPray answers questions over live SYSPRO data with read-only SQL or e.net queries it shows you, under the operator's own SYSPRO security, so a planner or accountant can ask the question in plain language and see exactly which tables or business objects produced the answer. Any write-back, such as posting a transaction, goes through a human approval step. The model is open-weight and runs on your own GPUs, so company data is not sent to a public AI service. ## Faq ### How does SYSPRO use a forecast in MRP? **Answer:** MRP reads forecast quantities by stock code and period alongside sales orders. Orders consume forecast inside a window so only the unconsumed forecast and extra orders create demand, and the run then suggests purchase orders and jobs. ### Which forecasting method should I use? **Answer:** Use a simple moving average or exponential smoothing for steady demand, a seasonal method where there is a calendar pattern, and planner or customer input for lumpy demand. Start simple and measure accuracy before adding complexity. ### Do I need forecasts for make-to-order items? **Answer:** Often not, because actual orders drive the plan. Forecast is most useful for stocked items and long-lead components where you must buy or build before the order arrives. ### Why is MRP planning more than I need? **Answer:** Likely the forecast is not being consumed, or it is too high or stale. Check the consumption window, the forecast quantities and old forecasts left on discontinued items. ## Search terms syspro forecasting, syspro forecast mrp, syspro demand forecast, syspro forecast consumption, syspro mrp forecast, syspro planning, forecast sales orders consumption, syspro statistical forecast --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json