Direct material systems: turning a planning signal into a supplier commitment

The signal originates in the plan, not in an inbox

Direct material demand comes out of the production plan rather than from a requisition somebody raises. If the chain from orders and forecast, through the planning run, to a supplier release is broken anywhere, a human patches it and the patch becomes permanent. The symptoms are easy to spot: buyers keeping private spreadsheets of what is really due, planners emailing shortages instead of trusting exception messages, and an order book that no longer matches the schedule. The system's contribution is making derived requirements visible and actionable, including the messages telling a buyer to pull a delivery in or push it out.

Agreements with releases versus discrete orders

For repeat material, raising a discrete order per delivery is administratively expensive and commercially weaker than an agreement with releases called against it. The agreement holds price, terms and horizon; the release states what to ship and when. Dividing that horizon into a firm zone the supplier may build to and a forecast zone they may plan against but not produce is the mechanism that shares risk explicitly rather than by argument after the event. Candidate systems vary in how well they support this, and a plant running repetitive supply on discrete orders is usually manufacturing avoidable work for both parties.

Confirmation is where the promise diverges

A release states what you want. What you have is what the supplier confirmed. If acknowledgements arrive by email and never reach the record, planning keeps working from requested dates nobody agreed to, and the shortage arrives as a surprise. Capturing confirmations in the system — with the supplier's own date and quantity, not a blanket accept — shifts planning from optimistic to real. It also generates the data to identify suppliers who habitually confirm late or short, which is more actionable than a delivery performance score computed long after the material was needed.

Receipt, inspection and matching inside a plant

Goods arrive against a release, not against an invoice, and the matching rules must cope with partial deliveries, over-shipment tolerance, unit-of-measure conversion and material physically present but not yet released by quality. Receiving into a quarantine status that planning can see but cannot allocate keeps the plan honest. The recurring failure is a receipt booked straight to available stock so the paperwork keeps moving while the pallet sits in an inspection cage, which produces a schedule built on material that may not lawfully be used.

The item parameters that quietly govern everything

A direct material system is only as good as its parameters. Lead times entered when the part was introduced and never revisited, minimum quantities copied from an obsolete price list, order multiples that ignore how the supplier actually packs, and safety settings nobody owns will all generate releases that suppliers ignore or that build stock nobody wanted. Give each parameter a named owning role, review the ones with the largest effect on a defined cycle, and treat a supplier consistently shipping something other than what was released as a data problem before assuming it is a behaviour problem.

Frequently asked questions

Can direct and indirect buying run in one system?

They can share a platform but should not share a process. Indirect buying is request-driven, approval-heavy and largely one-off. Direct buying is plan-driven, repetitive and schedule-based, and pushing it through a requisition and approval workflow adds delay without adding control. Where a single platform is used, configure two distinct flows, and resist pressure to make production material follow the office approval path, because buyers will simply raise orders outside it.

Who should own supplier lead times in the item master?

Purchasing owns the value because they negotiate it, but it needs a periodic reality check against actual receipt performance, which the system can produce automatically. The failure pattern is a lead time set optimistically at introduction, never challenged, and quietly compensated for by planners adding their own buffer. That double buffer inflates stock while still producing shortages. Make the planned-versus-achieved comparison a routine review item with a named owner rather than an occasional project.

What goes wrong first when planning parameters go stale?

Trust goes first. Planners notice that exception messages are wrong more often than right, start working from their own list, and stop actioning what the system produces. Once that happens the data decays further because nobody is correcting it, and the eventual verdict is that the software was inadequate. Recovery takes a deliberate cleanse of the parameters driving the most volume, plus visible evidence to the planning team that the output is now worth reading.