Blog / MES
MES · Guide

MES vs APS: The Schedule and What Really Runs

OL By OEE Lab Editorial |Updated 2 November 2026 |6 min read
Short answer

An APS (advanced planning and scheduling) tool works out the best order to run jobs, given finite machines, changeovers, labor and material. An MES sends that work to the floor and records what actually happens. The APS decides what should happen, and the MES reports what did, so the next schedule starts from the truth.

A schedule is a prediction. It stays right until the first breakdown, the first late pallet or the first order that runs slower than its standard.

The split between APS and MES follows from that: one is good at building the prediction, the other at knowing when it has gone wrong.

What an APS does best

An APS takes open orders and fits them to real constraints: machine capacity, setup times, tools, crews, material arrival and due dates. Then it searches for a good sequence, a thing no person can do well across hundreds of jobs.

The sequence matters most on lines with changeovers that depend on the previous product. Here is a small example.

From / toBCD
A206040
B-6015
C50-25
D3020-
Example changeover times in minutes, from the product on the left to the product on top. The line starts on A.

Starting from A, the other three products can run in six orders, with total setup anywhere from 55 to 130 minutes. Here are the worst, a middle case and the best.

A, D, B, C (40 + 30 + 60)130 min
A, B, C, D (20 + 60 + 25)105 min
A, B, D, C (20 + 15 + 20)55 min
Same products, same quantities. The best of the six orders saves 75 minutes against the worst.

A planner can spot this with four products. With 40 orders on 6 machines and due dates, it takes a solver.

Fast changeovers help every schedule, as SMED changeover reduction shows, and changeover frequency covers the trade-off with batch size.

What an MES does best

The MES takes the next job to the machine, confirms the material and the setup, then counts the output. When the line stops, it records for how long and why.

It answers the question that the APS cannot answer alone: where are we now? The order may be 40% done, the machine may be down, or the lot may be on hold.

The loop that keeps a schedule alive

ERP to APS: orders, due dates, BOMs, stock and routings.

APS to MES: the sequenced job list per machine, with planned start and end times and the frozen part of the schedule.

MES to APS: job started, quantity done, quantity left, machine state, scrap and the real run rate.

The cycle time of this loop matters. If the APS only learns what happened from a nightly upload, it plans the next day from yesterday's fiction.

On re-planning, avoid overreacting. Many plants freeze the next few hours so a small delay does not reshuffle the whole floor, and let a supervisor approve larger moves.

Standard rates versus demonstrated rates

Many schedule misses trace back to one input: the time an order is expected to take. An APS is only as honest as its rates.

Example: an order for 20,000 good pieces on a line with an ideal rate of 100 a minute. Over the last month, the MES shows this product really delivers 70 good pieces per minute of scheduled run time.

Planned at ideal: 20,000 / 100 = 200 min
Demonstrated: 20,000 / 70 = 286 min
Gap: 286 - 200 = 86 min per order
Every order planned at the ideal rate looks fine until the next job starts 86 minutes late.

Those 70 pieces a minute reflect stops, slow running and scrap, the same losses that OEE measures. So OEE data by product is a direct input to good planning, and OEE on mixed products shows how to keep it fair.

Take care with what you count. If the APS already models changeovers and planned breaks, leave them out of the rate, or they get counted twice.

Also keep the ideal rate and the planning rate as separate numbers. OEE must keep measuring against the ideal, while the schedule uses what you can currently deliver.

Detailed scheduling: the contested border

This border is contested because detailed scheduling is on the MES side of many function lists. The Object Management Group's MES breakdown, influenced by the MESA model, includes operations and detail scheduling, and many MES products offer it.

Scheduling in an MESDedicated APS
Typical horizonShift to a few daysDays to months
Typical methodRules, drag and dropConstraint-based search or optimization
ScopeLine or areaPlant or multi-plant, with what-if plans
Fits whenFew constraints, daily re-sequencingMany constraints and tight capacity
Typical, and products vary. Check what each one really does with your own orders.

The other overlap is who may change the sequence on the floor. If the supervisor swaps two jobs in the MES, the APS needs to hear about it, or the two lists will disagree by lunchtime.

A sensible order of work

1
Measure real rates
Collect output per product and line for at least a few weeks.
2
Find the bottleneck
Use your data to see which machine limits the plant.
3
Schedule the bottleneck first
Sequence it by hand or by simple rules, then let the rest follow.
4
Then consider an APS
If the sequence still loses time or misses dates, an optimizer now has honest inputs.
Each step gives the next one better data.

You can stop at any step. A few products on dedicated lines can run on a production board and takt time.

With steady demand and a simple sequence, a list from the ERP and an MES to dispatch it may be all you need. High mix, shared bottlenecks and due date pressure are what justify an APS.

An APS with no MES can work if confirmations reach the ERP quickly, though you will fight late data. An optimizer fed with bad rates just gives bad plans faster.

Bottleneck analysis helps with step two, and plan attainment gives you the score that shows whether the schedule is improving.

How a schedule loses the floor

  • Using nameplate speed for planning. You will be late, every day, by the amount of your losses.
  • Ignoring setup matrices. One average setup time hides the choice shown in the example above.
  • Replanning every five minutes. Operators stop believing the list, and the schedule loses its value.
  • Hiding overruns. If orders are closed early to look good, the rates feeding the APS are wrong.
See it live in Fabrico

A good schedule needs real rates. Fabrico is a light MES for OEE: stops, speed and counts arrive live from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, so you can see what each line actually delivers per shift before you plan on it. See how Fabrico's light MES for OEE works.

What people ask next

What is the difference between MRP and APS?

MRP calculates material needs and normally assumes unlimited capacity. APS adds finite capacity and sequencing. Many plants use both, and the MRP side is covered in MES vs MRP.

Can an MES do scheduling without an APS?

Many can, with dispatch lists, rules and drag-and-drop boards. That is enough for short horizons and few constraints. The MES modules page shows where this fits in the function list.

What data does an APS need from the shop floor?

Job status, quantity done and left, machine state, real run rates and scrap, as close to live as you can get them. A setup time matrix is also key, and the changeover time guide shows how to measure it.

Does better scheduling improve OEE?

It can reduce changeover time and waiting for material, which raises availability if those stops count in your planned time. It will not fix breakdowns or slow running. See how to count changeovers in OEE.

How often should an APS replan?

A common pattern is a full plan daily or per shift, with small updates when something big happens. Freeze the next few hours so the floor can prepare material and tooling.

See OEE per line and shift on your own machines

Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.

Book a Fabrico demo

← More OEE Lab guides · SMED changeover reduction →

OEE essentials: OEE calculator · What is OEE · Six big losses · Availability · TEEP · Cost of downtime · OEE benchmark · Free OEE tracker · Free OEE software · How to track downtime · How to implement OEE · Light MES vs full MES