CPG Scheduler
All posts

Rough-Cut Capacity Planning for CPG: Pressure-Test the Master Schedule Before It Hits the Floor

schedulingcpg operationsforecastingoeesequencing

Rough-Cut Capacity Planning for CPG: Pressure-Test the Master Schedule Before It Hits the Floor

The plan that can't be built

Every planner has seen it: a clean, confident master production schedule that quietly overloads the one CIP-limited filler in week 34, or stacks three promotional SKUs onto the same case-packer during your seasonal peak. The forecast looks great. The math on the demand side is airtight. And then the floor can't build it — so you scramble, expedite, cut runs short, and blow ship dates you already promised.

The problem usually isn't demand accuracy. It's that nobody asked the simple question first: does this plan physically fit on our lines, tanks, fillers, and labor?

That's what rough-cut capacity planning (RCCP) is for. It's the cheap, fast feasibility check that catches an unbuildable plan before you commit materials through MRP and before you turn a forecast into a customer promise. This post walks through what RCCP is, where it lives, the four classic techniques, and a concrete CPG worked example you can copy.

What RCCP is — and where it lives in the planning stack

Rough-cut capacity planning is a mid-range check that validates your master production schedule (MPS) against available resources — machines, labor, and critical materials — before that MPS becomes the driver for MRP. If the plan isn't feasible, you either add capacity or revise the MPS, and you usually iterate a few times until load and capacity reconcile.

RCCP is a layer in the classic MRP II hierarchy that Oliver Wight and others built on top of basic material requirements planning. The ladder looks like this:

  1. S&OP — sets the volume by family.
  2. Master scheduling — turns the plan into a buildable SKU-level MPS.
  3. Rough-cut capacity planning (RCCP) — validates the MPS against critical resources.
  4. MRP — explodes the MPS into component and material requirements.
  5. Capacity requirements planning (CRP) — the finer, work-center-level check after MRP.
  6. Detailed scheduling / execution — sequencing on the floor.

The key distinction: RCCP validates the master plan against a handful of critical resources; CRP validates the detailed plan after MRP at full work-center resolution. RCCP is deliberately coarse. It's a fast screen, not an execution schedule.

RCCP also connects tightly to concepts you may already run. It answers whether a plan is buildable, while time fences govern when you're allowed to change it. It's how you surface the constraint at the planning horizon before drum-buffer-rope schedules around it on the floor. And it confirms that a leveled Heijunka or EPEI wheel actually fits inside available hours.

The core idea: load vs. available capacity

RCCP compares two numbers, bucket by bucket (usually weekly):

  • Load — the capacity the MPS demands from a resource.
  • Available capacity — what the plant can realistically deliver.

The word "realistically" matters. There are two capacity numbers you must never confuse:

  • Design capacity — the theoretical maximum output, the nameplate number.
  • Effective capacity — the lower, achievable output after real constraints: quality losses, delays, material handling, changeovers, and sanitation.

RCCP must load against effective capacity. If you plan against nameplate speed, every week looks fine on paper and every week disappoints on the floor.

Remember too that maximum plant output equals the output of the slowest step — the bottleneck sets the ceiling. You can temporarily push a bottleneck above 100% with overtime, deferred preventive maintenance, or over-speeding a machine, but those boosts are short-lived by nature. Plan around the constraint; don't assume you can heroically run it hot forever.

CPG adds its own wrinkles to "available capacity":

  • Shared resources — one filler feeding a dozen SKUs, so load stacks even when each SKU looks small.
  • CIP and sanitation windows — hours that are simply unavailable for production.
  • Allergen and sequence-dependent changeovers — see sequence-dependent changeover scheduling and the hidden cost of changeovers for why these hours are real and non-trivial.
  • Tank and holding constraints — blend or hold capacity that caps upstream throughput.

The building block: the bill of resources

RCCP runs on a bill of resources (BOR) — a list of the resources (labor and key equipment) needed to make a saleable product. It complements the bill of materials (BOM): the BOM tells you what goes into the product; the BOR tells you what it takes to make it.

The BOR's job is to highlight critical resources — the ones in short supply or with long lead times. You don't model every conveyor and every operator. You model the constraints. For most CPG plants that's a short list: the filler, maybe a labeler or case-packer, a blend tank, and a labor pool.

A simple BOR expressed as hours per 1,000 cases might look like this:

Resource Hours per 1,000 cases
Blend tank 1.8
Filler (bottleneck) 4.0
Labeler 2.2
Case-packer 1.5
Direct labor 9.5

Multiply the MPS quantities by these factors and you get the load on each critical resource. That's the whole engine.

The four RCCP techniques, simplest to most detailed

There are four classic techniques. They all start from the MPS but differ in data requirements and precision. Pick the lightest one that still exposes your true constraint.

1. Capacity Planning Using Overall Factors (CPOF). The fastest and least data-hungry. It uses overall historical ratios — for example, total plant labor hours per 1,000 cases — and allocates that across resources by historical percentages. Good for a first-pass sanity check when your BOR data is thin.

2. Capacity Bills. Uses the BOR to translate the MPS directly into resource hours by work center. More accurate than CPOF because it reflects the actual resource mix each SKU consumes. This is the sweet spot for most CPG operations.

3. Resource Profiles. Capacity bills plus time-phasing. It offsets for lead time so the load lands in the correct week rather than lumping it all into the ship week. Useful when your critical resource sits several days upstream of finished goods.

4. Capacity Requirements Planning (CRP). The finest technique, and technically the next layer down — it runs after MRP at the work-order level using routing, lead time, and inventory status. Not RCCP proper, but worth knowing as the detailed check RCCP hands off to.

Which technique, when? Start with CPOF if you're building the discipline from scratch. Move to capacity bills once you have a trustworthy BOR for your critical resources. Add resource profiles only if lead-time offset is misplacing your load across weeks.

A worked CPG example

Let's make it concrete. Suppose your bottleneck is a single filler with an effective capacity of 80 hours per week after sanitation, changeovers, and PM are subtracted from the 120 running hours the line is staffed for. Your MPS for three SKUs over four weeks (in thousands of cases) is:

SKU Wk 1 Wk 2 Wk 3 Wk 4
A 8 8 12 6
B 5 6 6 5
C 4 4 7 4

The filler consumes 4.0 hours per 1,000 cases (from the BOR). Convert the MPS to load hours:

Week Total cases (000s) Filler load (hrs) Available (hrs) Status
1 17 68 80 OK (85%)
2 18 72 80 OK (90%)
3 25 100 80 Overload (125%)
4 15 60 80 OK (75%)

Week 3 is the problem: 100 hours of demand against 80 hours of effective capacity. RCCP just earned its keep — you found this in a spreadsheet, not on the floor at 2 a.m.

Now walk the fix options, cheapest first:

  • Level/pull ahead. Weeks 1, 2, and 4 all have slack. Pulling the SKU A spike or part of SKU C forward into weeks 1–2 smooths the load below 80 — classic production leveling. Check your time fence rules before you shuffle frozen-zone orders.
  • Add a shift or overtime in week 3. Temporarily lift effective capacity — but remember over-100% boosts are short-lived and erode into deferred PM.
  • Offload to an alternate line. If a second filler can run SKU C, route it there for the week.
  • Revise the MPS. If none of the above works, the plan isn't feasible and the master scheduler adjusts the promise — before MRP fires.

That iteration — flag, reconcile, re-check — is RCCP working exactly as designed.

What to do when RCCP says "no"

Overloads get fixed on the demand side or the supply side.

Demand-side: move a promotion out of the peak week, enforce available-to-promise (ATP) discipline so sales can't book against phantom capacity, or reshape the order pattern. See stop firefighting stockouts for the flip side — how honest capacity signals prevent the reactive scramble.

Supply-side: overtime, an added shift, alternate routing, or outsourcing.

For seasonal CPG peaks, it helps to name your capacity strategy up front:

  • Lead — add capacity ahead of demand (build for the peak early).
  • Lag — add only after demand materializes (conservative, risks shortfall).
  • Match — add incrementally as demand grows.
  • Adjustment — flex with overtime and temporary shifts around a stable base.

Most importantly, RCCP gives sales and operations one honest number to argue from. The perennial tension between what sales wants to sell and what the plant can build gets a lot less heated when both sides are looking at the same load-vs-capacity chart.

Limits and pitfalls — keep it honest

RCCP is powerful precisely because it's simple, but simplicity has costs:

  • It's an infinite-capacity screen. RCCP flags overloads; it does not sequence work. You still need finite scheduling and drum-buffer-rope on the floor to decide run order.
  • A leveled MPS can still hide bottlenecks. Lot-sizing decisions at lower BOM levels can create short-term capacity spikes that a smooth top-level MPS never reveals. Large lot sizes both inflate inventory and create bottlenecks. RCCP must be supplemented by finer checks.
  • Garbage in, garbage out. Stale BOR data, ignoring changeover and sanitation time, or loading against design instead of effective capacity will make every RCCP output a comforting lie.

Running RCCP in a headless, agentic scheduler

Most plants still run RCCP as a quarterly spreadsheet exercise — which means it's stale the moment the forecast moves. In a headless, agentic scheduling setup, RCCP becomes a continuous, automated what-if check: re-run whenever the forecast shifts or a time fence rolls, and test scenarios in seconds — can we support growing this category 15%? What happens if we add a line, or lose a week to a planned shutdown? The check stops being a report and becomes a guardrail.

Key takeaways

  • Build a bill of resources for your critical resources only — the filler, the tank, the labor pool — not everything.
  • Always plan against effective capacity, net of sanitation, changeovers, and PM. Never nameplate.
  • Run the lightest technique that exposes your real constraint: CPOF to start, capacity bills for most operations, resource profiles when lead-time offset matters.
  • Reconcile every overload before releasing the MPS to MRP — level, add capacity, reroute, or revise the plan.
  • Re-check whenever the forecast changes or a time fence rolls. RCCP is a discipline, not a one-time report.

Sources

More from the journal