Does production planning come before scheduling?

Yes. Demand becomes a production plan, the plan becomes a master production schedule, the master schedule releases work orders through MRP, and the scheduler sequences those orders onto machines. Scheduling also feeds back: actuals from the floor should correct the schedule the same day and the plan at its next revision.

Can the same person do both planning and scheduling?

In smaller plants, usually yes, and it works because the feedback loop lives in one head. It becomes a problem at scale: planning needs weekly rhythm and forecast thinking, scheduling needs continuous attention to live floor status. When one person does both under pressure, scheduling wins the day and planning quietly stops happening.

What is the difference between a production plan and a master production schedule?

The production plan states volumes by product family per period, months out. The master production schedule breaks that into specific end items by week, checked against rough-cut capacity, and is the contract between sales and the plant. Below it, MRP releases the work and purchase orders that the shop schedule then sequences.

Why do plans and schedules disagree?

Because they age at different speeds and rarely share data. Plans are built on standard rates and forecasts that drift toward optimism; schedules are built from morning snapshots that go stale by mid-shift. Without floor actuals flowing back into both, the planner blames the floor, the floor blames the plan, and both are partly right.

Which metrics track planning vs scheduling performance?

For scheduling, schedule attainment: the share of scheduled jobs completed as scheduled in the period. For planning, adherence to plan and forecast accuracy at the family level. Tracked together for a few weeks, they show whether misses originate in the plan's assumptions or the schedule's execution, which ends the blame cycle with data.

How does an MES fit between planning and scheduling?

The MES layer is where the schedule meets the floor: dispatching, execution, and data collection. An AI-native MES like Harmony AI connects machines, software, and paperwork into one live record, so actuals flow back up the cascade automatically, keeping the schedule current during the shift and feeding the plan honest numbers at its next cycle, with no rip-and-replace of the ERP.

Production Scheduling vs Production Planning

Q: What is the difference between production planning and production scheduling?

Planning works in weeks to months at product-family level and answers what to make, how much, and with what capacity and materials. Scheduling works in hours to days and commits specific jobs to specific machines in a specific order. The plan sets targets and secures resources; the schedule pays those commitments out in machine time.

Planning decides what and how much over months; scheduling commits machines and hours today. Where the line sits, how a plan becomes a schedule, and why the feedback loop between them is the part that breaks.

Production planning decides what to make, how much, and with what capacity and materials, over weeks to months. Production scheduling decides which specific job runs on which machine, in what order, starting when, over hours to days. The plan sets the targets; the schedule commits the resources.

The two terms get used interchangeably in most plants, and the confusion is not free: when nobody is sure where planning ends and scheduling begins, capacity problems get discovered on the floor instead of in the plan, and floor problems get blamed on people instead of the plan that made them inevitable. This post draws the line precisely, shows how a plan becomes a schedule step by step, and covers where each discipline breaks. The deep dives on each half are production planning and what is production scheduling.

What is the difference between production planning and production scheduling?

They differ on four axes: horizon, granularity, question answered, and what it costs to be wrong.

	Production planning	Production scheduling
Horizon	Weeks to months, sometimes quarters	Hours to days, at most a couple of weeks
Granularity	Product families, volumes, aggregate capacity	Specific jobs, machines, operators, start times
Question	What, how much, with what resources?	Which job, which machine, what order, when?
Typical owner	Planner, supply chain, S&OP process	Scheduler, production supervisor
Main inputs	Forecast, orders, aggregate capacity, material lead times	Routings, run rates, machine calendars, changeovers, live status
Output	Production plan, master schedule, purchase plans	Dispatch list, sequence, start/finish times
Cost of error	Wrong capacity or material, discovered weeks later	Wrong sequence, discovered by lunchtime
Revision cycle	Weekly or monthly	Daily, ideally continuous

Planning and scheduling compared. The horizon difference drives everything else: planning can rely on averages, scheduling cannot.

The horizon difference is the deep one. Over a month, run rates and downtime average out, so a plan built on standards is fair. Over a shift, nothing averages out: this changeover ran long, this press is down now. Scheduling lives with variance the plan is allowed to ignore, which is why the two need different data and different revision speeds.

What does production planning cover?

Planning is the resource-commitment layer. It takes demand, from the forecast and the order book, and decides whether the plant can meet it: how much capacity to allocate by product family, what materials to buy and when, whether to build inventory ahead of a peak or add a shift. Its main artifacts are the production plan by family, and below it the master production schedule, which states what will be built by week. Rough-cut capacity planning sanity-checks those numbers against aggregate capacity. Planning's job is to make sure that when the schedule tries to sequence the work, the capacity and material actually exist.

What does production scheduling cover?

Scheduling is the resource-commitment layer's cashier: it pays out the plan in specific machines and hours. It takes the orders the plan released, applies routings, run rates, changeover rules, and machine calendars, and produces the sequence each work center runs, ideally finite, never loading a resource past its real capacity, a distinction covered in finite vs infinite scheduling explained. Where planning can revise weekly, scheduling revises whenever the floor changes, because its output is executed within hours of being written.

Demand cascades through planning into scheduling and onto the floor. The dashed return path is the part most plants are missing: without floor actuals flowing back, the plan and schedule both drift from reality.

How does a plan become a schedule?

The handoff is a narrowing, from families and weeks down to jobs and minutes. In order:

Demand becomes a plan. The forecast and order book are balanced against aggregate capacity and material lead times, producing volumes by product family per period.
The plan becomes a master schedule. Family volumes break into specific end items by week, checked against rough-cut capacity.
The master schedule releases orders. MRP explodes the bill of materials, purchase orders go out, and work orders with quantities and due dates drop into the scheduling window.
Orders become a sequence. The scheduler places each work order on a machine with a start time, respecting routings, changeovers, material arrival, and real capacity.
The sequence becomes a dispatch list. Each work center gets its ordered queue: run this, then this. This is the artifact a supervisor can execute without asking a question.
Actuals flow back. Completions, scrap, downtime, and delays return upstream, correcting the schedule today and the plan next cycle. Without this step the cascade is a one-way street to fiction.

Where do planning and scheduling break down?

They break in opposite directions. Planning breaks by optimism: standards that assume the plant of three years ago, forecasts treated as facts, and capacity checked against nameplate rates nobody has hit since installation. The failure surfaces weeks later as material shortages and impossible master schedules. Scheduling breaks by staleness: a sequence built at 6 a.m. from yesterday's paperwork meets its first breakdown by 8, and by 10 the supervisor is running the floor from memory. The failure surfaces the same day as expedites and overtime.

Planning fails by assuming capacity that is not there, and the bill arrives weeks later. Scheduling fails by freezing at 6 a.m. while reality keeps moving, and the bill arrives by lunch.

And each blames the other. The planner points at the floor for missing the schedule; the floor points at the plan for being unrunnable. Both are usually right, because the real defect is the missing feedback loop between them, the dashed line in the diagram above. Adherence to plan and schedule attainment are the two metrics that make the gap measurable instead of political.

What do the standards and data say?

Primary-source context for the boundary between the two:

The ANSI/ISA-95 standard (IEC 62264), maintained by the International Society of Automation, draws this exact line architecturally: enterprise-level planning activities sit at Level 4, while detailed scheduling, dispatching, and execution sit in manufacturing operations management at Level 3, closer to the floor and on faster cycles.
The ASCM/APICS body of knowledge separates the planning hierarchy, sales and operations planning, master scheduling, MRP, from shop-floor control, the scheduling and dispatching layer that commits specific resources.
The scale is real: the Bureau of Labor Statistics counts roughly 12 to 13 million U.S. manufacturing jobs, and in most of those plants the plan and the schedule are still reconciled by hand, morning by morning.

How does Harmony AI connect the plan to the floor?

The broken part of the cascade is almost never the planning math or the sequencing logic; it is the feedback path. Plans are built on standards, schedules are built on snapshots, and the floor's reality reaches both too late. Harmony AI attacks exactly that path. As an AI-native MES, Harmony AI connects machines, existing software, and paperwork into one live operational record, so actuals flow up the cascade automatically: production counts, downtime, changeover times, and material status reach the schedule as they happen and the plan as it revises. AI agents act on the gap, re-sequencing around a down machine, flagging orders the miss puts at risk, and drafting the reporting that used to consume a planner's morning.

There is no rip-and-replace: the ERP keeps planning, the planner keeps deciding, and Harmony AI deploys in person, on your floor, connecting what you already run. That is the pattern in the CLS case study, where paper production records became real-time visibility and automated reporting against one source of truth. See how Harmony AI works, or start smaller: baseline your own gap with two weeks of schedule attainment data and our calculators and tools.