Short answer: planning and scheduling are two different jobs that get sold as one. Planning decides what to make and order, and roughly when, working in weeks and days. Scheduling decides when a specific job runs, on which machine, in what order, working in hours and minutes. Planning is mostly the MRP calculation on materials; scheduling is about capacity and sequence. Confusing them is why companies buy planning software and still miss delivery dates: they solved the material question and left the hours question untouched.
Planning and scheduling: the division of labour
| Planning | Scheduling | |
|---|---|---|
| Answers | What do we make and order, roughly when? | When does this job run, on which machine, in what order? |
| Unit of time | Weeks and days | Hours and minutes |
| Constraint | Materials and lead times | Capacity, setups, sequence |
| Output | Purchase and manufacturing orders | A sequence per work centre |
| Breaks when | Bills of materials or lead times are wrong | Routing and setup times are wrong |
You need them in that order. Scheduling a plan with no materials behind it produces a beautiful sequence for jobs that cannot start. But planning without scheduling gives you order dates and no idea whether the shop can actually deliver them, which is exactly where the trouble usually is.
Finite versus infinite capacity
The single most useful distinction in this market, and the one vendors are vaguest about.
- Infinite capacity scheduling assumes a work centre absorbs whatever you throw at it. Jobs get placed on the date they are needed, regardless of whether the hours exist. Simple, fast, and honest enough when you have slack.
- Finite capacity scheduling respects the real hours per work centre. When capacity runs out, work moves forward or backward and the promise date changes. Harder, slower, and the only version that tells you the truth once something is a bottleneck.
Most ERP systems schedule finitely against work centre capacity. Most spreadsheets schedule infinitely, which is why the spreadsheet always says the date is fine.
The check most planning articles skip
Before comparing tools, check whether the hours exist at all. Enter one work centre and see.
Demand
What has to come off the line.
Capacity
What the work centre can actually deliver.
One work centre, one week, indicative. Real scheduling weighs every work centre, sequence, changeovers and priorities at once. What this does show is the thing planning software cannot fix: if the hours are not there, the date is not real.
Notice the band boundaries, and note that they are not arbitrary. Somewhere around 85% utilisation, queues start growing sharply, and past roughly 95% they explode. This is queueing behaviour, not pessimism: as a resource approaches full load, every disturbance (a breakdown, a rush order, an operator off sick) has nowhere to go, so it turns straight into delay.
Which leads to the most useful sentence on this page: planning a bottleneck at 100% utilisation is planning to be late. Software cannot fix that. Only capacity, outsourcing, or a different promise date can.
Our opinion: most scheduling problems are measurement problems
We will be direct, because this is where the money is wasted.
When a manufacturer tells us their scheduling is a mess, we ask one question first: when did you last measure a routing time? The answer is usually a variation of “those came over from the old system” or “someone estimated them at go-live”. Everything downstream then inherits that guess. The schedule looks precise to the minute and is wrong by hours, and no tool can out-calculate wrong input.
Three positions we will defend:
- Setup times are the most underestimated number in manufacturing. They rarely appear in routings at full length, and in small-batch production they can outweigh run time entirely. If your schedule ignores changeovers, it is fiction with a Gantt chart on top.
- A perfect schedule that nobody follows is worth less than a rough one everybody follows. Shop-floor adoption beats algorithmic quality almost every time. If the plan lives on a screen the operators do not look at, you do not have a schedule, you have an opinion.
- If you are firefighting daily, more scheduling detail will not help. Finer planning on an overloaded shop produces a more precise account of why you are late. Fix the load first.
And the one that costs us: a large share of “we need scheduling software” conversations should end in “measure your routings and setups for a quarter first”. That is unglamorous, it does not need a purchase order, and it fixes more delivery reliability than most implementations do.
Choosing production scheduling software
Once you genuinely need a tool, these are what matter:
- Finite capacity, honestly implemented. Ask specifically how the tool behaves when capacity runs out. Does it move the date, or quietly overload the work centre?
- Setup and sequence handling. Can it model sequence-dependent setups (colour changes, material changes) where the order of jobs changes total setup time? This is the main reason to look beyond an ERP scheduler.
- Rescheduling behaviour. Real shops change hourly. How fast and how disruptively does the plan respond, and can planners lock or protect jobs?
- Shop-floor visibility. The schedule has to be visible where the work happens, or it is not real. This is where scheduling meets MES.
- Integration with your ERP. Scheduling needs routings, orders and stock. Separate tools need those synchronised, which is a project of its own.
- Planner trust. Can a planner see why the tool put a job there, and override it without breaking everything? An optimiser nobody trusts gets bypassed within weeks.
When your ERP already covers it
Be honest here, because it saves real money. If your ERP schedules manufacturing orders against work centre capacity, shows load per work centre, handles routing and setup times, and reschedules when things move, you have what most mid-sized manufacturers need.
Separate APS (Advanced Planning and Scheduling) software earns its place with sequence-dependent setups, many shared and constrained resources, or genuine optimisation across scenarios. That is a real category with real value, and it is a smaller group of companies than the marketing suggests.
Planning and scheduling in Odoo
In Odoo, both layers sit in Manufacturing on the same database as sales, purchasing and inventory:
- Material planning through the MRP run and reordering rules, covered in depth on the MRP system page.
- Work centres with capacity, including operating hours, efficiency and time before and after production.
- Routings with operation and setup times, so a manufacturing order occupies real hours rather than an assumption.
- Scheduling against capacity with a planning view and work centre load, so overload becomes visible instead of arriving as a surprise.
- Rescheduling when orders move, with the consequences flowing through to delivery dates.
- Execution feedback via Shop Floor, so the actual times land back in the system and your next schedule uses measured numbers instead of estimates.
That last point matters more than any scheduling feature: a closed loop between plan and reality is what makes routings accurate over time. That is also why we keep returning to the same advice, in job costing as much as here: measure, then plan.
Frequently asked questions
What is the difference between production planning and scheduling? Planning decides what to make and order and roughly when (weeks and days); scheduling decides when a job runs, on which machine and in what order (hours and minutes).
What is finite versus infinite capacity scheduling? Infinite ignores available hours and places jobs on the needed date; finite respects real capacity and moves work when hours run out. Finite is what you need once something is a bottleneck.
Do I need separate production scheduling software? Often not. A modern ERP scheduling against work centre capacity covers most mid-sized manufacturers. APS earns its place with sequence-dependent setups and many constrained resources.
Why do delivery dates slip even with planning software? Mostly utilisation and measurement: queues grow sharply above about 85% load, and routing and setup times are often estimates rather than measurements.
Can Odoo do production planning and scheduling? Yes, via MRP for materials and work centre capacity scheduling with routings, setup times and a planning view. For sequence-dependent optimisation, an APS tool alongside Odoo remains stronger.
Not sure whether your problem is scheduling, capacity, or measurement? Book a free Odoo scan - we look at your routings, setup times and work centre load, and tell you honestly which of the three you are dealing with.
Read more: MRP system explained · What is a MES? · MES vs MRP vs ERP · Manufacturing software: the categories · Odoo for manufacturers · The TARGET method