Advanced Planning and Scheduling (APS): The Benefits, and Why ERP Scheduling Keeps Slipping

Most factories already have a scheduling function. It is in the ERP, and it produces a plan that the shop floor treats as a suggestion. That gap — between the plan on the screen and the sequence people actually run — is what advanced planning and scheduling exists to close.

It is worth being precise about why the gap exists, because it is structural rather than a configuration mistake.

Infinite capacity is the whole problem

Classic ERP and MRP planning works backwards from a due date using fixed lead times and assumes capacity is available. It answers when work should start if nothing is in the way.

Something is always in the way. Two orders want the same machine at the same hour. The tool needed is on another job. The operator qualified for that process is on the other shift. MRP does not see any of it, so it produces a plan that is arithmetically correct and physically impossible, and the floor quietly reorders the work into something that can actually run.

Once that happens twice, the plan stops being a commitment and becomes a document. Everything downstream degrades with it — promised dates, purchasing timing, staffing — because they were all derived from a schedule nobody follows.

APS plans against finite capacity: real machines, real tooling, real qualified people, real changeover times, in a real sequence. The output is a plan that can be executed, which is a different thing from a plan that adds up.

The benefits that actually appear, and roughly when

Quotable dates you can hit. This is the first one to show and usually the most valuable commercially, because sales stops padding lead times defensively and the padding was costing orders.

Less changeover time, without buying anything. Sequencing by tooling, colour, material or temperature rather than by due date alone recovers hours a week on lines where setup is significant. It is the closest thing to free capacity in a factory.

Lower work in progress. Finite-capacity plans release work when the constraint can take it rather than as early as possible, so material stops piling up between stations. Cash and floor space both improve, and quality often does too because defects are found closer to their cause.

Faster, calmer response to disruption. A machine goes down at 06:00 and the question is what it costs and what moves. With APS that is a rescheduling run. Without it, it is three experienced people and a morning.

And a quieter one that matters more than it sounds: the argument moves. When the plan is credible, discussion shifts from whose list is right to what to do about the constraint.

Why APS on top of a blind shop floor underdelivers

A scheduling engine is only as good as its picture of the present. Give it yesterday’s completion data and it will optimise a factory that no longer exists, confidently.

This is the most common reason APS disappoints. The engine is fine. It is being fed shift-end data entry, so by mid-morning the schedule describes a state that has drifted, the floor notices, and the plan loses authority exactly the way the MRP plan did.

APS on live MES data is a different product from the same APS on assumptions. Sequence recalculates when a machine actually stops, not when someone types that it stopped. If the shop floor is not yet reporting in something close to real time, connectivity is the prerequisite, not an enhancement.

The data you need before an APS project is worth starting

Routings that match reality, including the operations people actually perform rather than the ones documented in 2014. Setup and changeover times that reflect the current tooling, and are sequence-dependent where that is true — light-to-dark, thin-to-thick.

A machine and resource model with real constraints: which equipment can run which product, what tooling is shared, which qualifications a station requires. Shift calendars including planned maintenance windows, breaks and holidays. And material availability that is a live figure rather than a monthly count.

None of this needs to be perfect. It needs to be better than the planner’s mental model, because that mental model is the incumbent and it is surprisingly good. A schedule the floor trusts less than the experienced planner’s instinct will not be followed, whatever it cost.

APS and supply chain planning are not the same horizon

They get conflated because both are planning. APS schedules inside the plant, at the level of operations and hours, over days and weeks. Supply chain planning coordinates across sites and suppliers, at the level of quantities and dates, over weeks and months.

They need to share numbers and they solve different problems. A supply plan that is not grounded in achievable plant schedules produces promises the factory cannot keep; a plant schedule with no view of supplier lead times sequences work for material that will not arrive. Running both against one live plan is the point — sales, purchasing and production disagreeing productively about one set of dates rather than privately about three.