The engine works. We say that with confidence because we've watched it run high-volume food & beverage production, complex multi-level discrete manufacturing, and multi-site distribution networks. When MRP "doesn't work" in Sage X3, the problem is almost never the engine — it's the setup: BOMs that don't reflect how production actually runs, planning parameters left at defaults, lead times nobody validated, and demand signals the system was never given.
The result is always the same: planners stop trusting the suggestions, planning moves to spreadsheets, and the ERP quietly stops being the system of record. We've restored MRP adoption in operations that had run on spreadsheets for years — in one food & beverage distributor's case, in 60 days. See case studies →
What correct Sage X3 MRP setup actually involves
- BOM and routing accuracy. MRP is only as honest as the product structures it explodes. That means real quantities and yields, scrap factors that match the floor, and routings whose lead times reflect actual throughput — not the optimistic numbers from the original implementation.
- Phantom structures. Complex products often carry intermediate assemblies that exist logically but never sit in stock. Configured correctly, phantom BOMs explode straight through to raw material requirements without generating phantom work orders or stranded inventory levels. Configured wrong, they generate noise that buries the real suggestions.
- Co-products and byproducts. Process and food manufacturing breaks naive MRP assumptions: a recipe's output isn't one item. Take a 1,000-gallon batch recipe — protein, vegetables, herbs, packaging — where trimming produces a byproduct that flows back into stock as planned supply for another product. Sage X3 models this with byproduct lines on the BOM, so the planning run nets both sides: demand for the primary output, supply of the byproduct. Getting the yield percentages and replenishment logic right is meticulous work — and it's exactly the work that separates "MRP can't handle food production" from MRP running the plant.
- Planning horizons and time fences. Demand inside the firm horizon behaves differently from demand at week 12. Horizons, time fences, and reorder policies (MRP, reorder point, period coverage) have to match each item's real procurement and production rhythm — produce with 3-day lead times and proteins with 3-week lead times do not plan the same way.
- Demand quality. MRP nets against the demand you feed it. Forecast management, sales order hygiene, safety stock logic, and inter-site demand all shape whether suggestions are credible. A perfect engine fed garbage demand produces confident garbage.
- Suggestion management discipline. Planners need a working rhythm: run the calculation, review by exception, firm and convert. If the suggestion workbench shows 4,000 lines because parameters were never tuned, nobody reviews anything — and the spreadsheet returns.
When companies call us about MRP
- MRP was never turned on. The original implementation went live on finance and inventory; planning stayed in Excel "temporarily," and temporary became years.
- MRP generates suggestions nobody trusts. Quantities are wrong, dates are wrong, or the volume of noise makes review impossible.
- The operation outgrew the original setup. New product lines, co-pack arrangements, multi-site sourcing, or volumes the original parameters never anticipated.
- MPS was never implemented. Everything plans straight from raw demand, so the schedule whipsaws with every order change and the floor stops believing the plan.
How an MRP recovery engagement runs
We start with a structured assessment: BOM and routing audit, parameter review across a representative item sample, a test planning run compared against what planners actually did, and interviews with the people running the spreadsheets — because the spreadsheets encode the business rules the system is missing. From there: corrected structures and parameters, a controlled re-launch on a pilot item family, side-by-side validation until the suggestions earn trust, then rollout and planner training. Adoption is the deliverable — not a configured module.
Frequently asked questions
What's the difference between MRP and MPS in Sage X3?
MRP calculates material and production requirements from all demand sources. MPS is a stabilizing layer for selected key items: it locks a master schedule inside the planning fence so MRP plans components against a committed plan instead of reacting to every demand fluctuation. Most mid-market manufacturers benefit from MPS on finished goods with volatile demand or constrained capacity.
Why is Sage X3 MRP not generating the suggestions we expect?
The usual suspects: missing or wrong reorder policies on items, demand not visible to the run (forecast not loaded, order status excluded), BOM or routing gaps, planning horizon cut short, or site/sequencing setup excluding locations from the calculation. We maintain a working troubleshooting guide.
Can Sage X3 MRP handle co-products, byproducts, and recycling loops?
Yes — byproduct lines on BOMs model secondary outputs as planned supply, and co-product structures split yields across outputs. Recycling loops (byproduct re-consumed upstream) are the hardest case and need careful structure design to avoid circular planning references — it's advanced configuration, and it's exactly where most setups give up and planners go back to spreadsheets.
How long does it take to fix MRP in Sage X3?
A focused recovery on a live system typically runs 60–90 days from assessment to validated adoption on the first item families, depending on BOM cleanup volume and data quality. The 60-day adoption restoration in our case study is representative of a well-scoped engagement.
Do we need new software or add-ons to make planning work?
Almost never. In our experience, mid-market manufacturers rarely need more software — they need the platform they already own configured by people who understand both Sage X3 and how a plant actually runs.