Material Requirement Planning (MRP) Systems in Apparel ERP

Material Requirement Planning answers two questions for every material on every open order: what quantity must be ordered, and when must it be ordered. It answers them by calculation, not by judgement. The system takes the order quantity, explodes the bill of materials into gross requirements, nets those requirements against stock on hand and material already scheduled to arrive, and produces a net requirement. It then offsets that requirement backwards by supplier lead time to give an order release date. The whole method fits in one paragraph, and factories that run it honestly rarely run out of material.

The alternative is the spreadsheet habit: a planner estimates fabric from experience, adds a generous buffer, and orders. That habit works until it does not, and it fails in both directions at once. Fabric gets over-ordered because the buffer is applied to stock already held. Trims run short because nobody exploded the trims BOM with the same care as the fabric BOM. MRP fixes both failures with the same short calculation, applied to every BOM item without exception.

The equation is: net requirement equals gross requirement minus on-hand stock minus scheduled receipts. Gross requirement comes from exploding the BOM: order quantity multiplied by consumption per garment, broken down by size and colour where the style needs it. On-hand stock is what the warehouse actually holds of that exact item. Scheduled receipts are quantities already ordered with promised arrival dates. The order of operations after netting matters as much as the netting itself: the wastage allowance is added after netting, and lot sizing is applied last.

Take an order of 12,000 polo shirts with a fabric consumption of 1.65 metres per shirt. Gross requirement is 12,000 multiplied by 1.65, which is 19,800 metres. On-hand stock is 4,500 metres and scheduled receipts are 6,000 metres, so net requirement is 19,800 minus 4,500 minus 6,000, which is 9,300 metres. A 3 percent wastage allowance added after netting gives 9,300 multiplied by 1.03, which is 9,579 metres to order. The same logic covers buttons: 12,000 shirts multiplied by 5 buttons gives a gross requirement of 60,000 buttons; with 15,000 on hand and no scheduled receipts, net requirement is 45,000 buttons, rounded up to supplier lots of 10,000, which gives 5 lots totalling 50,000 buttons.

Adding wastage before netting is the classic error. Applied to gross requirement, a 3 percent allowance on 19,800 metres would add 594 metres, most of it wastage on fabric the factory already holds. Applied after netting, the same 3 percent adds 279 metres, which is the honest buffer on material that must actually be bought and processed. The difference, 315 metres on one order, multiplied across a season of orders, is real money sitting in a warehouse as excess stock.

Knowing what to order is half the plan. Knowing when to order it is the other half, and MRP sets the date by counting backwards. Start from the production start date, the day cutting must begin, and subtract the supplier lead time plus a safety margin. If production starts on day 40 and fabric lead time is 25 days, the order must be released by day 15. If the release date has already passed when MRP runs, the system raises a past-due exception, which is the most urgent message in the whole run.

Safety margins deserve a policy, not a habit. A planner who adds a week to every lead time creates a warehouse full of early arrivals and a false sense of security. A planner who adds nothing discovers that one late truck stops a line. The honest policy sets the margin from supplier reliability history: reliable suppliers earn a small margin, unreliable suppliers earn a larger one, and the margin is reviewed when reliability changes. Lead-time offsetting also exposes the orders that cannot be saved by expediting: when the required date falls before the earliest possible supply date, the choice is to negotiate faster supply or move production start, and that choice must be made the week MRP raises it, not the week cutting is due to start.

An MRP run produces two outputs: suggested order quantities, and exception messages. Planners who read quantities first and exceptions never are the reason MRP gets blamed for stockouts it actually predicted. The correct reading order is exceptions first. Past-due orders mean a release date has passed with material unordered: order immediately or expedite, escalated the same day. Shortage warnings mean net requirement exceeds available supply even after planned receipts: bring a receipt forward, split the order across suppliers, or replan production. Excess stock means holdings exceed open-order requirements: hold new orders and check whether the BOM or the buffer is wrong. Lead-time breaches mean the required date is unreachable: negotiate or move production.

Every override of an exception message must be logged with a reason and an owner. Silent overrides are how the same stockout repeats every season with a different excuse. The weekly planning meeting should review overrides before it reviews new orders, because overrides are where judgement lives, and judgement that is never reviewed never improves. A factory that clears its exceptions weekly will rarely be surprised by a shortage. A factory that ignores them will be surprised regularly, and will blame the system for predictions it chose not to read.

MRP trusts its inputs completely, which is both its strength and its danger. A BOM error propagates to every order using that style. A stale stock record, where the warehouse holds less than the system claims, produces a confident under-order. Unrecorded consumption, such as fabric issued without a system entry, creates the same gap. When MRP and the warehouse disagree, trust the warehouse count and fix the record; planning on a wrong stock figure will confidently order the wrong quantity.

The defence is a short routine: version every BOM change and regenerate MRP immediately after a change, reconcile stock before the weekly run, and treat any large or surprising suggestion as a signal to check inputs before ordering. MRP is not a black box to be obeyed. It is a calculation to be checked, and the check takes minutes: gross requirement, stock, receipts, net, buffer, lot size, release date. A planner who can walk that chain for any material on any order understands the system. A planner who cannot is operating a machine whose failures will always arrive as surprises.

Run MRP weekly, because buyers change quantities and colour ratios far more often than once a month. A monthly run plans for a factory that no longer exists. Each weekly cycle follows the same order: reconcile stock and scheduled receipts, regenerate the plan on current BOM versions, read exceptions first and clear them with named actions, release orders by their offset dates, and log every override. The cycle takes a morning for a medium factory, and it replaces a week of spreadsheet reconciliation and emergency expediting.

Over a season, the rhythm compounds. Over-ordering falls because wastage stops being applied to stock already held. Trims shortages fall because trims run through the same netting as fabric. The planning meeting changes character: instead of arguing about quantities, the team clears exception messages and makes dated decisions. That is what a working MRP system looks like, not a dashboard, but a weekly habit of netting honestly, reading exceptions first, and ordering on the date the arithmetic demands.