S&OP Process Setup

Executive summary

S&OP process setup is the work of installing a monthly planning cycle in which sales, operations, and finance agree on one plan: what customers will want, what the plants and suppliers can deliver, what inventory will result, and what that does to cash and margin. Sales and operations planning is not a software category and not a meeting. It is a decision process with a fixed calendar, named owners, and a single set of numbers.

Most mid-market manufacturers and distributors already have the pieces. Sales has a forecast, often a hopeful one. Operations has a production schedule, built around what is on the floor this week. Finance has a budget that was set last year. Each is internally coherent and none of them agree. The gaps show up as expediting, overtime, excess stock in the wrong items, and late orders that the customer experiences as an unreliable supplier.

The fix is modest in design and hard in discipline. Hold one meeting per month with a fixed agenda. Review four things in order: last month's results, the demand plan, the supply plan, and the gaps between them. Decide the gaps in the meeting or assign a named owner and a date. Record what was decided. The company does not need a perfect forecast. It needs a visible, shared forecast and a habit of reacting to the gap between plan and reality early enough to do something about it.

This guide covers the cycle, the roles, the metrics (forecast accuracy and bias, plan attainment, inventory against policy), templates you can copy into a spreadsheet, a worked example with reconciled arithmetic, and a 90-day plan to get the first full cycle running.

When to use this

Use this guide when:

• Sales, operations, and finance each quote a different number for next quarter's volume.
• Expediting, premium freight, and weekend overtime are routine rather than exceptional.
• Inventory is too high in some items and short in others at the same time.
• On-time delivery is slipping and the root cause analysis keeps pointing to "we did not see it coming." See on-time delivery root cause attribution for the attribution side.
• You have a new owner, a sponsor, or a lender asking for a forecast you can defend.
• You are growing fast enough that the informal "ask the plant manager" method has stopped working.

Do not start with S&OP if you cannot produce a clean item master, a bill of materials, and a month of shipment history by product family. A planning cycle on dirty master data produces confident, wrong plans. Fix the data first, or choose a family level coarse enough that the data you have is reliable.

Diagnostic questions

Answer these before designing the cycle. They tell you how much of the process already exists and where the first meeting should focus.

1. Is there a single written demand forecast by product family for the next 12 months, and who owns it?
2. Does operations plan to that forecast, or to open orders and gut feel?
3. When sales and operations disagree, who decides, and where is the decision recorded?
4. How often does the production schedule change after it is released, and what are the top three reasons?
5. Is there a calendar of known events (promotions, customer shutdowns, tooling changes, supplier lead time changes) that feeds the plan?
6. Do you measure forecast accuracy and bias, and by what level (item, family, customer)?
7. What is the quoted lead time to customers, and how does it compare to the actual lead time over the last six months?
8. Which constraint (a machine, a skilled trade, a supplier, a shipping dock) most often limits output, and does the plan show it?
9. Does finance see the inventory and cash consequence of the operations plan before it is executed, or after?
10. How many people can describe the current plan in one sentence, and do they say the same sentence?

The framework: five steps, one calendar

The classic S&OP cycle has five steps. In a mid-market company the whole cycle fits inside one calendar month, with a pre-meeting working week and one executive decision meeting.

Two rules that keep the cycle honest

Unconstrained demand first. The demand plan states what customers are expected to want, not what the plant can make. If you cap demand at capacity before the supply review, you hide the gap that the meeting exists to find.

One set of numbers. The approved S&OP plan is the volume forecast that finance uses for the budget reforecast, that operations uses for the master schedule, and that sales uses for customer commitments. When any function maintains a private forecast alongside it, the process has failed quietly. See rolling forecast vs annual budget for how the approved plan feeds the financial forecast.

Step-by-step process

Step 1: Set the scope and the unit of planning

1. Choose the product families. Group items that share a routing, a bottleneck, or a customer pattern. Start with the three to six families that make up most of revenue. Owner: Head of Operations with Head of Sales. Output: a family list with a definition for each.
2. Choose the planning horizon and bucket. Monthly buckets for 12 to 18 months suit most mid-market companies. Use weekly buckets only for the first one to three months if the process is volatile. Owner: Planning lead. Output: a horizon and a bucket rule.
3. Define the time fences. A frozen window (no changes without executive approval), a slush window (changes allowed within limits), and a free window (open to revision). Set them from your actual lead times, not from a template. Owner: Head of Operations. Output: written fences.

Step 2: Build the baseline

4. Pull 24 months of shipments by family (or as much history as is reliable), plus open orders and backlog. Owner: Planning lead. Output: a history table.
5. Pull the constraints: capacity by work center for the bottleneck, supplier lead times for long-lead materials, and labor availability for skilled trades. Owner: Head of Operations. Output: a constraint sheet.
6. Pull inventory by family against its policy (target, minimum, maximum). If no policy exists, use the safety stock approach in inventory optimization: turns, safety stock, and excess and the safety stock calculator. Owner: Planning lead. Output: inventory position.
7. Compute last quarter's forecast accuracy and bias with the formulas in the next section. Owner: Planning lead. Output: the baseline accuracy number you will improve on, not compare to someone else's.

Step 3: Run the first demand review

8. Statistical baseline plus judgment. Start from a simple baseline (for example, trailing average adjusted for trend and season). Sales adds known events and customer information on top, and writes each adjustment down with a reason and an owner. Owner: Head of Sales. Output: the demand plan.
9. Flag risks and opportunities separately from the plan. Do not bury a 50-percent-likely deal in the base number. List it, with a probability that sales commits to and a date by which it resolves. Owner: Head of Sales. Output: a risk and opportunity list.

Step 4: Run the first supply review

10. Load the demand plan against capacity at the constraint, by month. Show the percent of available constraint hours the plan consumes. Owner: Head of Operations. Output: a load profile.
11. List the gaps (demand exceeds supply) and the surpluses (supply exceeds demand). Quantify each in units and in revenue or cost. Owner: Planning lead. Output: a gap table.
12. Generate options for each gap: overtime, an added shift, outsourcing a step, a supplier expedite, a change in inventory build, a customer commitment change, or a price or allocation decision. Cost each with finance. Owner: Head of Operations and CFO. Output: an option table with costs.

Step 5: Hold the executive meeting

13. Chair it for decisions, not for review. Use the agenda below. Every open gap leaves the room with a decision, or with a named owner and a date. Owner: CEO or GM.
14. Record the decision log. One line per decision: the decision, the reason, the owner, the due date, and the expected effect. Owner: Planning lead. Output: the log, circulated within one working day.

Metrics and KPIs

The point of the metric set is to tell you whether the process is improving, not to rank people. Use three groups: how good the forecast is, how well the plan is executed, and what the plan costs.

Setting targets without invented benchmarks. Take the trailing best of your own accuracy by family as the first-year goal and measure improvement against your own baseline. Where a family is naturally lumpy (project orders, a single large customer), measure at a higher level of aggregation or by customer segment, and do not compare it to a steady-demand family. If you want an outside reference, use a named public peer's disclosed service levels or a published industry body, and cite it. Do not quote a percentage from a blog and assign it to your plant.

Templates

Demand and supply plan by family (copy into a spreadsheet)

Assumption log

Executive S&OP agenda (60 to 90 minutes)

1. Decision log from last month: done, late, dropped (10 min).
2. Performance review: forecast accuracy and bias, plan attainment, delivery to commit, inventory versus policy (10 min).
3. Demand plan: changes since last month and the risk and opportunity list (10 min).
4. Supply plan: constraint load, gaps, and options with cost (20 min).
5. Financial view: revenue, margin, inventory, and cash under the proposed plan versus the budget (10 min).
6. Decisions: approve one plan; assign each open item an owner and a date (15 min).
7. Next month: calendar, topics, and any changes to the process itself (5 min).

Decision log

Worked example (illustrative)

The figures below are made up to show the method. They are not benchmarks.

A manufacturer of one product family plans six months ahead. Standard capacity is 1,150 units per month at the constraint. Opening inventory is 400 units. The inventory minimum is 450 units. The approved demand plan is 1,000, 1,100, 1,250, and 1,300 units for the next four months.

Projected ending inventory = prior inventory + production − demand.

Total demand across the four months is 4,650 units against standard capacity of 4,600 (4 × 1,150). That is a 50-unit shortfall in the horizon. Month 4 ends 100 units below the minimum, because the minimum is 450 and the projected inventory is 350.

The plant has not missed a shipment yet. The S&OP view finds the problem three months before it appears. Options the supply review might cost with finance:

• Run overtime to add 100 units in month 3 or month 4, bringing month 4 back to 450.
• Build ahead in month 1 or 2 by adding units before the constraint saturates.
• Ask sales whether the month 4 demand is firm, or whether part of it can be shifted by a week or two.

The decision belongs to the executive meeting because the options trade cost against service. Whichever is chosen, it is written to the decision log.

Forecast accuracy and bias on the same numbers

Suppose the forecast and actuals for four past months were:

WAPE = 300 ÷ 4,600 = 6.5%. Bias = 0 ÷ 4,600 = 0%.

The bias is zero because over- and under-forecasting cancel out over the four months. That is why bias alone can look healthy while the plant lurches between too much and too little. Reported alone, a zero bias would hide the 6.5% error. Reported alone, a 6.5% error would not tell you whether the problem leans one way. Use both, and look at them by family, because a blended number hides the family that is driving the miss.

Common mistakes

1. Treating S&OP as a forecasting project. The forecast is an input. The product of the process is a decision about gaps. Companies that spend six months on the statistical model and never hold the executive meeting have not set up S&OP.
2. Capping demand at capacity. If the demand plan is already constrained, the meeting has nothing to decide.
3. Planning at item level. Item-level forecasts are noisy and take hours to review. Plan at family level and let planners disaggregate.
4. Letting each function keep a private plan. The budget, the master schedule, and the sales target must trace to the approved plan.
5. No decision log. Without a record of what was decided and what happened, each meeting restarts from memory, and the same arguments recur.

AI prompt assistance

AI can prepare the analysis and tidy the documents. It cannot know your constraints or your customers. Supply the data, and check the output against your own records before anything enters the plan.

1. Compute accuracy and bias from your own history.

I will paste a table of monthly forecast and actual shipments by product family for the last 12 months. Compute WAPE and bias for each family and for the total. Show each calculation step. List the two families that contribute the most absolute error and say whether each leans high or low.

Check: recompute one family by hand, and confirm that the actuals match your ERP shipment report.

2. Draft the assumption log from sales notes.

Here are my sales team's notes on expected changes for the next six months. Convert each note into a row with: family, month, change in units, reason, owner, a probability if one was stated, and a date it will be resolved. Mark any row where a number or an owner is missing instead of guessing.

Check: send the table back to the Head of Sales and confirm that every figure is theirs, not the model's.

3. Identify gaps from a plan table.

Here is a demand plan, a supply plan, opening inventory, and an inventory minimum by month for one family. Compute projected ending inventory for each month, list every month below the minimum, and quantify the shortfall in units. Do not suggest solutions yet.

Check: verify the arithmetic on two months, and confirm that the capacity figure is the constraint's actual available capacity, not the nameplate figure.

Implementation guide

Downloadable assets

The demand and supply plan table, the assumption log, the executive agenda, and the decision log above are designed to be copied into a spreadsheet. The safety stock calculator can help set the inventory minimum you test the plan against.

Related resources

• Operations guides: the full operations library, including capacity, quality, and visual management.
• Capacity utilization and throughput analysis: finding the constraint your supply plan has to respect.
• Inventory optimization: turns, safety stock, and excess: setting the inventory policy the plan is tested against.
• On-time delivery root cause attribution and the OTD root cause calculator: measuring whether the plan is improving delivery.
• Margin expansion playbook: where planning discipline shows up in margin.

Frequently asked questions

What is S&OP in simple terms?

S&OP is a monthly process in which sales, operations, and finance agree on one plan for demand, supply, inventory, and cash. It ends with an approved plan and a list of decisions with owners and dates. It is a decision cycle, not a forecasting tool or a software purchase, and it works with spreadsheets if the discipline is there.

How long does S&OP process setup take?

A basic cycle can run within 90 days: about a month to define families, roles, and the calendar, then two or three monthly cycles to settle the routine. Data cleanup often takes longer than the meetings. Expect the first cycle to be rough, and judge progress by whether decisions are made and tracked.

Who should attend the executive S&OP meeting?

The CEO or general manager chairs, with the heads of sales, operations, and finance, and the planning lead who prepares the data. Add supply chain or quality leaders where they own a constraint. Keep the group small enough that decisions are made in the room rather than deferred to another meeting.

What is the difference between forecast accuracy and bias?

Accuracy measures how large the forecast errors are, while bias measures whether they lean in one direction. A forecast that overshoots in some months and undershoots in others can show zero bias and still have large errors. Track both, and look at them by product family rather than only in total.

Do we need software to run S&OP?

No. Many mid-market companies start with a spreadsheet for the demand and supply plan, an assumption log, and a decision log, and add planning software later if volume or complexity justifies it. The binding constraints are clean master data, a fixed calendar, and a chair who forces decisions.

How this works in a System of Performance

An S&OP cycle only improves if what was assumed, what was decided, and what happened stay connected from one month to the next. In most companies the forecast sits in one spreadsheet, the schedule in the ERP, the decisions in meeting notes, and the outcomes nowhere in particular. A Performis routine can take on a recurring job like this, with the scope and fit agreed with you: it works from the records your team supplies, prepares the work for your named reviewer, and keeps each approval on record. If you want to see where your own plan and results already disagree, bring it to us as a launch partner.