How do S&OP, the master schedule and detailed scheduling fit together?

S&OP reconciles demand, supply and financial plans at an aggregate level, commonly by product family. The master production schedule states what quantities of specific end items are planned for specific periods. Material requirements planning derives component needs from that schedule, while capacity planning checks the resources needed to make it. Detailed scheduling then assigns work to actual resources and controls its execution. A feasible family total does not prove that every item mix or due date is feasible.

Why can an MRP calculation call for an order when stock is on hand?

MRP compares time-phased gross requirements with projected available inventory and scheduled receipts. Suppose 20 units are on hand, 30 are due before week 4, and 60 are needed in week 4. With no safety stock and lot-for-lot ordering, the week 4 shortage is 60 minus 20 minus 30, or 10 units. A planned receipt of 10 is needed in week 4. If lead time is two weeks, its planned release belongs in week 2.

Does a higher inventory target always improve customer service?

Extra stock can reduce shortages when demand or replenishment varies, but its value depends on where the uncertainty occurs and which item the customer needs. Stock of an unrelated component cannot cover a finished-item shortage. Safety stock also raises holding cost and may become obsolete. Set a service objective, estimate demand and lead-time variation, place the buffer at a useful point, and compare the expected service gain with the added inventory investment.

CPIM Study Guide: Planning and Inventory Management

Start with concept 1 Choose a topic

Planning decisions connect across time and across functions. A sales and operations plan sets a workable direction for product families; a master schedule turns that direction into commitments for specific items; material, capacity and detailed schedules test whether those commitments can be met. Inventory can protect service, but it also ties up cash and can conceal problems. This guide follows those connections through the nine areas of Certified in Planning and Inventory Management (CPIM). Each concept starts with a decision a planner might face and uses a fully specified example to show the calculation or reasoning. Pay attention to what each plan assumes, which resources constrain it, and when new information should change it. Those distinctions make the methods useful beyond a single set of numbers.

Align the Supply Chain to Support the Business Strategy

1. Read the competitive setting

Competitive analysis becomes useful when it connects an external change to a capability you can improve. Start with the customer segment and the reasons buyers consider, choose or leave a supplier. Then examine rivals' prices, service, quality and capacity, alongside your own performance. A five-force analysis helps identify pressures such as rivalry and buyer bargaining power; a SWOT analysis separates external opportunities and threats from internal strengths and weaknesses. Neither framework makes the decision for you. Check customer evidence against operational measures, then identify the performance gap that matters most to orders and margin. A rival's price cut may warrant a pricing response, but persistent late delivery points first to promise accuracy, available stock or fulfillment flow. The current CPIM content outline includes competitive analysis within business strategy.

Worked example: A rival reduces its standard-item price by 5%. In the same month, eight of ten customers who explain a lost order cite late delivery; two cite price. Delivery reliability is the immediate operations priority because it addresses the more frequently reported loss. Check promised dates against actual dispatch and receipt dates to locate the delay. A 5% price match alone leaves those eight complaints unresolved.

Watch the distinction: Treating the most visible competitor move as the main customer problem can misdirect resources. Compare rival actions with customer evidence and your own measured performance before choosing a capability to improve.

Reference: [1]

2. Qualifiers, winners and customer value

An order qualifier is a minimum condition for being considered. An order winner is the attribute that makes a qualified offer preferable in a particular buying decision. The distinction depends on the market: a feature that wins one segment may only qualify a supplier in another. Establish the buyer's stated requirements first, then compare the offers that pass them. To find where operations creates the winning value, follow the order through entry, confirmation, inventory allocation, production if needed, picking and delivery. Measure elapsed time at each step and separate processing from waiting. A faster delivery promise may depend less on faster transport than on earlier order visibility or reliable component availability. Reassess qualifiers and winners as competitors improve, since yesterday's advantage can become today's minimum expectation.

Worked example: A buyer considers only certified suppliers that can deliver within five days. Two suppliers meet both conditions, but one offers replenishment in two days and wins. Certification and the five-day limit are qualifiers; the two-day service is the winner. If order entry takes two days and picking takes one, reducing order-entry delay offers a clearer route to the faster promise than shortening picking alone.

Watch the distinction: Calling every valued feature an order winner hides the decision rule. A buyer may value certified quality while treating it as a required entry condition rather than the reason for choosing one qualified supplier.

Reference: [1]

3. Choose the push-pull boundary

The push-pull boundary is the point where work changes from being driven by a forecast to being triggered by a specific order. Make-to-stock places that point after finished-goods production and suits predictable demand with short customer lead times. Make-to-order starts production after the order and suits greater variety when customers can wait. Assemble-to-order holds common modules in advance, then performs final configuration after the order. Compare the customer's acceptable lead time with the time needed for each remaining activity, including assembly, testing and delivery. Also compare the inventory risk of holding finished variants with the risk of holding shared modules. Postponing differentiation can pool demand across options, but only if components and final assembly capacity are available when orders arrive.

Worked example: Four common bases and ten color options create 4 × 10 = 40 finished combinations. Customers allow three days; final assembly takes one day and, hypothetically, delivery takes one more. Stocking the four bases and color modules permits assembly after the order and delivery in two days. Holding all 40 finished variants would tie inventory to uncertain color choices without improving the stated promise.

Watch the distinction: Assuming postponement eliminates inventory is tempting. It moves inventory toward shared components and unfinished units; shortages of an option module or assembly capacity can still delay the customer's configured order.

Reference: [1]

4. Make or buy with the constraint in view

A make-buy comparison begins with costs that change because of the decision. For internal production, include variable cost, avoidable setup and any opportunity cost of using scarce capacity. For purchasing, include the quote, inbound freight, inspection and other relevant acquisition costs. Exclude costs that remain under both choices. Then test feasibility: can the supplier meet the required specification, quantity and date, and can internal production do the same? Capacity matters especially at a constraint. Buying can free constraint hours for a product with greater contribution, even if the purchase price exceeds internal variable cost. Conversely, dependence on one supplier may warrant a buffer or qualification expense. State the decision horizon because a temporary purchase and a permanent transfer can have different effects on skills, fixed costs and supply risk.

Worked example: For 2,000 units, internal variable cost is 2,000 × $18 = $36,000, plus an avoidable $8,000 setup, totaling $44,000. A qualified supplier charges 2,000 × $21 = $42,000, assuming no added freight or inspection cost. Buying saves $2,000 on these inputs. Confirm the supplier's date and quality; any valuable alternative use for released internal capacity strengthens the buying case.

Watch the distinction: Comparing the $21 supplier price with only the $18 internal variable cost omits the avoidable setup. Including unavoidable allocated overhead would create the opposite distortion because that cost remains under either choice.

Reference: [1]

5. Match process and layout to volume

Process choice follows the combination of volume, variety and routing stability. A job shop groups similar equipment by function, allowing different products to follow different routes. That flexibility usually brings more movement, scheduling effort and queue time. A product-focused line arranges steps in production sequence. It can deliver steady flow and lower handling cost when volume supports dedicated resources and work is sufficiently repeatable. Batch production and cellular layouts sit between those ends: a cell groups resources for a family with similar routes. Consider setup time, required skills, equipment investment and expected demand before moving along this product-process pattern. A product's life cycle can change the economics as designs settle and volume grows, but age alone is not a reason to install a line. Test the proposed layout against actual flow and demand.

Worked example: A prototype shop makes 20 variants in lots of two, or 40 units across frequently changing routes. Flexible machines grouped by function suit that work. Later, one stable design sells 5,000 identical units each month. A dedicated line may then reduce handling and queues enough to justify its investment. Compare expected savings with the line's cost before changing the layout.

Watch the distinction: A high-volume label does not automatically justify an assembly line. Unstable designs, uneven demand or a constrained operation can leave expensive dedicated equipment underused despite an impressive annual volume.

Reference: [1]

6. Measure strategy and sustainability

A strategy needs measures that show both its intended gain and its side effects. Financial measures reveal cost and cash effects; operational measures show whether promises are met; environmental and social measures expose other consequences of the same choice. Define each measure's unit, time period and denominator so alternatives are comparable. For example, freight cost per order can fall while late orders and emissions per order rise. Convert rates into counts or totals at the expected volume, then compare the results with the organization's priorities and constraints. A balanced scorecard can make these relationships visible, but it does not turn unlike measures into one objective. When a decision improves one measure and harms another, identify the trade-off explicitly and determine whether a different route, carrier or service tier can improve the combination.

Worked example: For 1,000 orders, a transport change saves ($12 - $10) × 1,000 = $2,000. On-time delivery falls from 98% to 90%, so expected late orders rise from 20 to 100, an increase of 80. Emissions rise from 2 to 3 kg per order, adding 1,000 kg. The saving must be weighed against those service and environmental effects.

Watch the distinction: Calling the cheaper transport option an improvement because freight fell ignores the measures tied to customer and sustainability goals. Use the same order volume to translate all three changes into comparable business consequences.

Reference: [1]

7. Map and mitigate supply risk

Supply risk analysis starts with the actual network, including important suppliers behind direct suppliers. Map where critical materials, tools and processes originate, then look for common nodes such as a shared fabricator, port or information system. Estimate the effect of losing each node using recovery time, available inventory, substitute capability and customer commitments. Choose a response based on impact and risk tolerance. Prevention may reduce the chance of failure; mitigation limits its effect through inventory, alternate capacity or product redesign; recovery restores flow after an event. A backup is useful only when it is qualified and does not depend on the same vulnerable node. For a related treatment of exposure and response choices, see the risk management study guide. Review the map when suppliers or materials change.

Worked example: Two direct suppliers each offer 500 chips a week, apparently covering a 600-chip weekly need. Both buy wafers from one fabricator. If that fabricator stops, neither direct supplier can deliver, so the apparent 1,000-chip capacity provides no independent backup. Qualifying a second fabricator addresses the common node; shifting volume between the existing two suppliers does not.

Watch the distinction: Splitting orders between two supplier names does not necessarily diversify risk. Trace critical materials upstream before crediting a second supplier as protection against a disruption.

Reference: [1]

Conduct Sales and Operations Planning (S&OP) to Support Strategy

8. Place S&OP in the planning hierarchy

Sales and operations planning connects business direction with executable plans at an aggregate level. It typically considers product families, resource groups, inventory, backlog and money over a horizon long enough to change capacity or sourcing. The approved family volumes guide the master schedule, which translates them into specific end items and periods; material and detailed schedules then specify components and operations. Each level answers a different question. A feasible quarterly family total does not prove every week or model is feasible. Before disaggregation, check seasonal demand, family mix, available workdays, critical resources and supplier limits in each period. If the detail cannot support the aggregate plan, send that constraint back for reconciliation. The hierarchy works through repeated feedback, not a one-way release of numbers.

Worked example: A quarterly family supply plan is 12,000 units. An even split gives 12,000 ÷ 3 = 4,000 units per month. If the stated seasonal shares are 20%, 30% and 50%, the monthly volumes are 2,400, 3,600 and 6,000; they sum to 12,000. Check whether the third month's resources can make 6,000 before assigning item-level quantities.

Watch the distinction: Dividing a quarterly plan equally by three can hide a seasonal capacity gap. Preserve the approved total, then test the timing and mix against monthly resources before creating the master schedule.

Reference: [1]

9. Build the portfolio demand view

A portfolio demand view asks how many units customers will buy across the affected products, not how many each forecast reports in isolation. A new product can attract genuinely incremental customers, replace sales of an existing product, or do both. Begin with a baseline for existing products that excludes the launch effect. Estimate new-product demand and the portion transferred from products already in that baseline. Add only the incremental portion to portfolio demand, while adjusting the product mix for the transferred portion. Check whether the forecast also contains lost sales caused by retiring an old product or shortages during transition. Distinguish unit volume from revenue and contribution: a replacement sale may leave total units unchanged yet materially change margins or capacity load. Record the assumptions so actual sales can test them after launch.

Worked example: The established line's baseline is 1,000 units. A new model is forecast to sell 200, with 100 of those sales replacing established-line purchases. Expected established-line sales become 1,000 - 100 = 900. Portfolio demand is 900 + 200 = 1,100 units, or 1,000 + 200 - 100. Simply adding both forecasts would overstate total volume by 100 units.

Watch the distinction: Treating all new-product sales as incremental double counts customers who switch from an existing product. Estimate the transfer separately, then update both total demand and the product mix.

Reference: [1]

10. Balance aggregate supply and inventory

The aggregate inventory balance is a quick feasibility test: beginning inventory plus planned supply minus expected demand equals projected ending inventory. Keep all terms in the same unit and period, and include other known receipts, losses or backlog treatment if they matter. Compare the result with the approved ending target rather than assuming that a nonnegative balance is sufficient. A target may protect seasonal demand or a replenishment gap; it also ties up cash. When the projection misses the target, identify the size of the gap and the available levers. Supply can rise through overtime, outsourcing or an earlier receipt; demand may shift through accepted later dates or substitution. If none is attractive, management can explicitly change the inventory target or accept a service risk. The balance shows the gap, while capacity and cost analysis select the response.

Worked example: Beginning inventory is 200 units, planned production is 900 and demand is 1,000. Projected ending inventory is 200 + 900 - 1,000 = 100 units. Against a 150-unit approved target, the gap is 50 units. Produce 50 more, move 50 units of demand with customer agreement, or approve the lower ending balance after assessing its service risk.

Watch the distinction: A projected ending balance above zero does not mean the plan meets its inventory objective. Compare it with the target and explain the service or cash consequence of any difference.

Reference: [1]

11. Size a strategic capacity buffer

A capacity buffer is available capability above the expected load at a resource that could limit output. Calculate effective available hours for the period, including approved overtime but subtracting known maintenance and other downtime. Compare that capacity with the load implied by the proposed volume and mix. The remaining hours are the buffer against variation, urgent work or estimation error. A small positive buffer is not proof of robust performance if arrivals or processing times vary greatly. Also distinguish regular-time overload from total-capacity feasibility: overtime may make a plan possible, yet staffing and authorization still need to be in place. Locate buffers where they protect an actual constraint or important commitment. Spare hours at a nonconstraint cannot directly replace missing hours at a specialized bottleneck.

Worked example: A work center has 400 regular hours and 30 approved overtime hours next month, with no other downtime in the hypothetical plan. Expected load is 420 hours. Total capacity is 430, leaving a 430 - 420 = 10-hour buffer. The load exceeds regular capacity by 20 hours, so at least 20 overtime hours must actually be staffed.

Watch the distinction: Counting all 30 overtime hours as unused buffer overlooks the 20 hours already needed to cover regular-time overload. The remaining cushion is 10 hours, subject to the stated capacity assumptions.

Reference: [1]

12. Reconcile the financial plan

Financial reconciliation tests whether an operational response supports the business plan. For an incremental order, compare contribution that would be lost without the response with the additional cost required to serve it. Contribution should already deduct the normal variable costs used to produce and sell the units; subtract only costs added by the proposed response. Then check whether the choice changes other costs, such as expedited freight, quality loss or inventory holding. Cash timing and capacity commitments can matter even when the arithmetic looks favorable. Use the same horizon for revenue, cost and resource availability. A positive incremental contribution supports the response on the stated financial basis, but it does not establish feasibility if labor, materials or customer delivery dates remain unresolved. The management accounting study guide develops the contribution approach further.

Worked example: A 200-unit surge would yield $30 contribution per unit after normal variable costs. Unserved demand therefore puts 200 × $30 = $6,000 at risk. Overtime can make all 200 and adds $15 per unit, or $3,000. Assuming the sales are otherwise lost and no other costs change, accepting the overtime yields $6,000 - $3,000 = $3,000 incremental contribution.

Watch the distinction: Subtracting normal variable production cost again from an already calculated contribution understates the benefit. Check what the contribution figure includes, then deduct only costs created by the proposed response.

Reference: [1]

13. Approve a feasible volume and mix

When a resource is constrained, compare products by contribution per hour of that resource, subject to commitments and demand limits. First reserve capacity for orders that must be honored. Then calculate each remaining product's contribution divided by its constraint hours per unit. Allocate available hours to the highest value use until its demand limit is reached, then consider the next product. This is a short-run decision rule, so it assumes the stated contribution and constraint are relevant over the planning horizon. It also requires whole-unit quantities and any setup or minimum-lot limits to be checked. A product with greater contribution per unit can be less attractive if it consumes much more bottleneck time. After choosing a mix, reconcile the resulting units with material availability, other capacities and the approved financial plan.

Worked example: Product A contributes $40 per unit and uses two constraint hours, or $20 per hour. B contributes $20 and uses 0.5 hour, or $40 per hour. Of 500 available hours, a firm 100-unit A commitment uses 200. The remaining 300 hours make 600 B, assuming that demand exists. Total contribution is 100 × $40 + 600 × $20 = $16,000.

Watch the distinction: Ranking A above B because $40 exceeds $20 ignores the scarce hours each consumes. Reserve firm commitments, then compare contribution per constraint hour for the capacity still available.

Reference: [1]

Plan and Manage Demand

14. Separate independent and dependent demand

Independent demand comes from outside the product structure, such as customer orders or a forecast for a finished item. Dependent demand is calculated from a parent item's planned quantity and its bill of materials. Forecasting each component separately can conflict with the parent plan and obscure timing. To calculate a component's gross requirement, multiply the parent's planned release by the quantity of that component per parent, then aggregate requirements from all parents that use it. Next, account for usable on-hand stock, scheduled receipts, safety stock, yield and lot rules to find net requirements. Timing matters: a seal needed when pump assembly starts must be available by that start date, not merely by the pump's shipment date. Service-spare demand for the same seal may be independent and should be added separately.

Worked example: A forecast calls for 100 finished pumps, each requiring two seals. Gross dependent demand is 100 × 2 = 200 seals. If 50 usable seals are on hand, with no other demands, receipts, safety stock or losses, the shortage is 200 - 50 = 150 seals. The 100-pump forecast drives the calculation; it does not make the 200-seal requirement a separate forecast.

Watch the distinction: Calling every component requirement independent because the component is purchased confuses sourcing with demand origin. Its production requirement follows the parent plan, while separately sold replacement seals can have independent demand.

Reference: [1]

15. Adjust a forecast for a market event

A promotional forecast should isolate the event's incremental effect from sales moved between products or periods. Start with a baseline that excludes the promotion. Estimate the promoted item's added sales, then identify what share will come from other products in the same portfolio. Subtract those transfers when forecasting total portfolio units, while still adjusting the product mix and capacity load. A promotion may also pull later purchases forward, causing a post-event dip; account for that if the horizon spans both periods. Use comparable historical promotions, customer commitments or point-of-sale evidence where available, and note uncertainty because response may differ by channel. After the event, compare actual uplift and cannibalization with assumptions. Otherwise a successful promotion of one item can look like new demand even when total business barely changed.

Worked example: The portfolio baseline is 500 units for the promotion period. The promoted item is expected to gain 100 sales, but 70 are expected to replace purchases of another portfolio item. Total demand becomes 500 + 100 - 70 = 530 units. The promoted item's plan rises by 100; the other item's plan falls by 70. Net portfolio uplift is 30 units.

Watch the distinction: Adding all promoted-item sales to the portfolio baseline assumes every buyer is new. Separate transferred sales from incremental sales so supply planning does not overstate total volume.

Reference: [1]

16. Update with exponential smoothing

Simple exponential smoothing updates a prior forecast by a fraction of its latest error: next forecast equals prior forecast plus α times actual demand minus prior forecast. The smoothing constant α lies between zero and one for this method. A larger α responds more strongly to recent demand but can also react more to random noise; a smaller α changes slowly and may lag a real shift. Use consistent time buckets and units, and establish whether unusual events should be adjusted before updating. Simple smoothing is best suited to a fairly stable level. Persistent trend or seasonality calls for a method that models those patterns rather than expecting one smoothing constant to capture them. Track forecast errors after each update so responsiveness and bias can be judged from performance, not from the formula alone.

Worked example: Suppose the prior forecast for a week is 100 units, actual demand is 120 and α = 0.30. Error is 120 - 100 = 20 units. The next forecast is 100 + 0.30 × 20 = 106 units. It moves six units toward the latest actual; adopting 120 would treat one observation as the whole new demand level.

Watch the distinction: Using α = 0.30 to multiply actual demand alone gives 36, which discards the prior forecast. The update applies α to the forecast error, then adds that adjustment to the prior forecast.

Reference: [1]

17. Apply a seasonal demand pattern

A seasonal pattern distributes an annual or other aggregate forecast according to expected period shares. Confirm that the shares cover the same horizon and sum to 100%, then multiply each share by the aggregate forecast. The resulting period quantities should sum back to the total, apart from any explicit rounding adjustment. Seasonal shares describe timing, not necessarily growth: a larger third-quarter share does not by itself mean annual demand has risen. Examine whether the pattern is stable enough to use, especially after a product launch, channel change or major promotion. Next compare the timed forecast with capacity and inventory plans. If a peak period exceeds feasible output, earlier production and anticipation inventory may be needed. Keep demand timing distinct from supply timing; moving production earlier does not change when customers want the units.

Worked example: An annual forecast is 1,200 units, with quarterly shares of 20%, 30%, 35% and 15%. Multiplication gives 240, 360, 420 and 180 units. The shares total 100%, and 240 + 360 + 420 + 180 = 1,200. The third quarter has the largest demand, so capacity or inventory plans must support 420 units then.

Watch the distinction: Applying seasonal percentages without checking their sum can change the annual total unintentionally. Normalize or correct the shares first, then verify that the period forecasts add back to the approved aggregate.

Reference: [1]

18. Detect error and bias

Forecast error needs both a direction convention and a measure suited to the question. Define error as actual minus forecast: positive errors mean underforecasting, while negative errors mean overforecasting. Mean error or cumulative forecast error reveals directional tendency across periods, but positive and negative misses can cancel. Mean absolute deviation, or MAD, averages their absolute sizes and therefore shows typical miss magnitude in the original units. For products with different scales, percentage errors may help comparison, though near-zero actual demand can make them unstable. Review errors by item, location and horizon as well as in aggregate; opposing biases can cancel when combined. A period with zero mean error can still have costly shortages and excess inventory. Investigate recurring direction and size before changing the method or adding a buffer.

Worked example: With error defined as actual minus forecast, four errors are +10, -10, +20 and -20 units. Mean error is (10 - 10 + 20 - 20) ÷ 4 = 0. MAD is (10 + 10 + 20 + 20) ÷ 4 = 15 units. These four periods show no net directional error, but forecasts missed by 15 units on average.

Watch the distinction: A zero mean error does not mean accurate forecasts. Opposite misses cancel in the signed average; inspect MAD and the error sequence before concluding that demand is predictable.

Reference: [1]

19. Consume the forecast with orders

A forecast represents expected demand that has not yet become firm orders. Forecast consumption replaces the relevant forecast with customer orders as those orders arrive, preventing the same demand from appearing twice in the master schedule. The rule must specify the item, period and consumption window; orders in another period or for a different configuration may not consume this forecast. Within a simple same-period rule, remaining forecast is the greater of forecast minus firm orders and zero. Total demand is firm orders plus remaining forecast, equivalently the larger of orders and forecast. Once orders exceed the forecast, the excess is a genuine increase in planned demand. Monitor that change because material and capacity plans may need adjustment. The CPIM practice questions offer further scenarios for applying planning distinctions.

Worked example: A period forecast is 100 units. With 70 firm orders that consume it, 100 - 70 = 30 forecast units remain and total demand is 70 + 30 = 100. Another 50 firm orders bring orders to 120. Remaining forecast becomes zero, so total demand is 120 units, an increase of 20 above the original forecast.

Watch the distinction: Adding 70 orders to the untouched 100-unit forecast would plan 170 units for demand already represented in the forecast. Apply the stated consumption rule before totaling demand.

Reference: [1]

20. Shape demand within capacity

Demand management can change when or how customers request supply, but it must preserve a truthful record of what they still need. First compare requested volume with available capacity by period and identify any firm commitments that cannot move. Then consider actions such as offering later dates, substitutions, different service windows or incentives for off-peak demand. Test customer acceptance and the capacity of the receiving period before changing commitments. A deferred order becomes backlog or future-period demand; it is not a cancellation unless the customer actually cancels. Distinguish demand shaping from a supply response such as overtime, and compare the service and financial effects of each. If demand remains above capacity after feasible actions, prioritize according to commitments and communicate realistic dates. Record the reason for each move so forecast and service measures remain interpretable.

Worked example: A service team can complete 100 installations next week, while customers request 130. Thirty customers accept dates the following week, so next week's committed load is 130 - 30 = 100 jobs, matching capacity. The later week's plan gains 30 jobs. If that week already has 85 requests and capacity of 100, another 15-job gap remains and needs a further response.

Watch the distinction: Moving 30 jobs out of next week's schedule is not a reduction in total demand. Record their new due dates and check later capacity, or the apparent solution merely shifts a hidden overload.

Reference: [1]

Plan and Manage Internal Supply Sources

21. Project the master schedule

A master production schedule states when specific end items are expected to be completed. Its projected available balance (PAB) tests whether that schedule covers expected demand period by period. Start with the prior period's PAB, add the current period's master-scheduled receipts, and subtract the demand used for planning. Use one consistent rule for combining forecasts and customer orders so the same sale is not counted twice. Receipt timing matters: a unit completed after a customer's required date cannot cover that earlier requirement. A negative PAB exposes a timing or quantity problem, but a positive balance alone does not prove that materials and capacity can support the schedule. Compare the proposed schedule with those constraints, then revise receipts or demand commitments where necessary. The current CPIM content outline includes master scheduling within internal supply planning.

Worked example: Assume all receipts arrive before demand within their stated week. Opening finished-goods inventory is 80 units. Week 1 demand of 50 leaves a PAB of 80 - 50 = 30 units. A 100-unit master-scheduled receipt arrives in week 2 before demand of 60, so week 2 PAB is 30 + 100 - 60 = 70 units. Without that receipt, week 2 would be short by 30 units.

Watch the distinction: A week 2 receipt does not raise the week 1 balance. Adding all future receipts to opening stock hides shortages that occur before those receipts arrive.

Reference: [1]

22. Promise only available supply

Available-to-promise (ATP) identifies supply that has not already been committed to customer orders. For a discrete ATP calculation at a master-scheduled receipt, subtract firm orders due from that receipt through the period before the next scheduled receipt. The first ATP bucket may also include uncommitted opening inventory; later buckets normally use their own master-scheduled receipts. Forecasts help plan supply but are not firm customer commitments in this calculation. Check dates as well as quantities: a later receipt cannot support a promise due earlier. If ATP is insufficient, assess another receipt, a different date, or a feasible capacity change before confirming the order. Capable-to-promise goes further by checking whether material and capacity can create additional supply. ATP can be positive even when the physical stock has not yet arrived.

Worked example: A 100-unit master-scheduled receipt is due before the next delivery window. There is no uncommitted opening stock, and 65 units of firm orders are due before the following receipt. Discrete ATP is 100 - 65 = 35 units. A new 40-unit order exceeds that supply by 5 units. Promise at most 35 from this receipt and check another feasible date or source for the remaining 5.

Watch the distinction: Subtracting the forecast as well as firm orders can count the same demand twice. ATP concerns uncommitted supply; a separate feasibility check may still be needed.

Reference: [1]

23. Check rough-cut capacity

Rough-cut capacity planning tests a proposed master schedule against a small set of critical resources before detailed work-center scheduling. A bill of resources translates each planned end-item quantity into hours or other resource units. Sum the requirements by period and compare them with capacity actually available after maintenance, staffing limits, and other known losses. Use consistent units and the same time bucket for load and capacity. An overload is a planning exception: move production, add feasible capacity, change the mix, or revisit the master schedule. A spare hour at a different resource does not solve a bottleneck shortage unless work can genuinely move there. Rough-cut checks are approximate, so passing one is a reason to proceed to detailed planning, not a guarantee that every operation can be scheduled.

Worked example: A proposed weekly master schedule contains 300 units, each requiring 0.5 hour at a critical work center. Rough-cut load is 300 × 0.5 = 150 hours. The center has 140 available hours that week, leaving a 10-hour overload. If standards and available hours are valid, shift at least 20 units to another feasible period or secure 10 additional hours before approving the schedule.

Watch the distinction: Calling the schedule feasible because the factory has spare hours overall ignores where capacity is needed. The critical work center is short by 10 hours.

Reference: [1]

24. Explode the bill of materials

A bill of materials (BOM) expresses the parent-child structure of a product and the quantity of each component needed per parent. Explode requirements one level at a time: multiply the parent's planned order quantity by the quantity per assembly for each child. For a multilevel BOM, repeat that multiplication down the structure. Time matters alongside quantity. A child's gross requirement is generally tied to the planned release of its parent, because the component must be available when parent production starts. After calculating gross requirements, apply inventory records, scheduled receipts, lot-sizing rules, yields, and lead times to determine net requirements and releases. If the same component appears in several branches, sum its requirements across those branches. A wrong quantity-per assembly or obsolete revision can propagate a large error through the plan.

Worked example: Suppose each finished A requires two B assemblies, and each B requires three C parts. A planned release for 20 A therefore creates gross demand for 20 × 2 = 40 B. Those 40 B require 40 × 3 = 120 C. These are gross quantities. Any usable B or C on hand, scheduled receipts, or scrap allowances must be evaluated separately before ordering.

Watch the distinction: Subtracting C inventory before calculating all C requirements can miss demand from another BOM branch. Explode and combine gross needs first, then net the shared item.

Reference: [1]

25. Net and offset MRP requirements

Material requirements planning (MRP) converts time-phased gross requirements into planned receipts and releases. For each period, begin with projected available stock, add scheduled receipts, and subtract gross requirements. If the result falls below the required ending balance or safety stock, calculate a net requirement. A lot-for-lot rule makes the planned receipt equal that shortage; other lot rules may enlarge it. Offset the planned receipt backward by the item's lead time to find when its order must be released. The release, rather than the receipt, drives the gross requirements of lower-level components. Always verify receipt dates, unit of measure, BOM revision, and existing orders before acting on an MRP exception. A mathematically correct release date can still be impossible if it falls in the past or available capacity cannot meet the lead time.

Worked example: A component has a gross requirement of 60 units in week 4, with 20 usable units on hand and a scheduled receipt of 30 before week 4. Assume no intervening demand or safety stock. Available supply is 50, so the net requirement is 60 - 50 = 10 units. Lot-for-lot planning places a 10-unit receipt in week 4. With a two-week lead time, release it in week 2.

Watch the distinction: Scheduling the order release in week 4 confuses the need date with the start date. The two-week lead time requires a week 2 release.

Reference: [1]

26. Calculate work-center load

Capacity requirements planning turns orders into load at each work center using routing data. Multiply each order's quantity by its standard run time per unit, add any setup time charged to that center, and assign the hours to the periods when operations are expected to occur. Include released orders, planned orders, rework, and other known work so the comparison reflects the full queue. Compare load with available hours after staffing, calendars, and planned downtime; adjust for demonstrated efficiency only when the capacity measure has not already incorporated it. An overload calls for a feasible response such as rescheduling, splitting a lot, using an alternate routing, or adding shifts. Moving a job may solve one period's overload while creating another, so recheck the entire affected schedule.

Worked example: One work center is assigned 100 units of A at 0.25 run hour each and 80 units of B at 0.5 run hour each. Run load is 100 × 0.25 + 80 × 0.5 = 25 + 40 = 65 hours. With 60 available hours, it is overloaded by 5 hours. This comparison excludes any setup hours, which would increase the load.

Watch the distinction: A quantity total of 180 units says little about capacity without routing times. The two products consume different hours per unit at this work center.

Reference: [1]

27. Schedule final configuration

In assemble-to-order or configure-to-order operations, a master schedule may plan common modules before customers choose final options. The final assembly schedule translates actual configurations and due dates into specific assembly work and option-component needs. For each order, multiply the option quantity per finished unit by the order quantity, then check usable components, labor, and assembly capacity at the required dates. Reserving modules for one order reduces what remains for others, so the sequence of commitments matters. A component shortage may permit a partial release, a later date, or an approved substitution, but it should not be hidden by scheduling finished units that cannot be built. Compare actual completions with the final assembly schedule and update it when customer choices or incoming supply change.

Worked example: A customer orders 20 configured units, each needing two identical option modules. The order requires 20 × 2 = 40 modules, but only 30 usable modules are available before assembly. At most 30 ÷ 2 = 15 complete units can start from current stock. Ten more modules are needed for the other five units; promise their completion only after checking the replenishment and assembly dates.

Watch the distinction: Thirty modules do not support 20 finished units when each unit consumes two. Scheduling all 20 would leave five units waiting for missing parts.

Reference: [1]

Plan and Manage External Supply Sources

28. Choose a sourcing structure

Sole sourcing means only one supplier is available for a requirement, often because of a unique design, patent, or qualification constraint. Single sourcing means a buyer deliberately places all business with one supplier even though alternatives exist. Multisourcing allocates business to more than one qualified supplier. The choice affects unit price, coordination, learning, leverage, and continuity risk. A split may preserve two active sources, but its backup value depends on their independent capacity, tooling, and upstream dependencies. Assess whether each source can meet specifications and surge when another fails; a nominal second source with no available capacity offers little protection. Concentrating volume can improve economics, while maintaining an alternate can reduce exposure, so evaluate total cost and interruption consequences together.

Worked example: Two suppliers are qualified for a 1,000-unit monthly requirement. A 70%/30% allocation sends 1,000 × 0.70 = 700 units to the first and 300 to the second, making this a multisource arrangement. Ordering all 1,000 from either one would be single sourcing because another qualified supplier remains available. It would be sole sourcing only if no viable alternative existed.

Watch the distinction: Using one supplier this month does not prove sole sourcing. The distinction turns on whether another viable supplier exists, not how many purchase orders were issued.

Reference: [1]

29. Compare total acquisition cost

A quoted unit price is only one part of acquisition cost. Bring bids to the same quantity, currency, delivery point, and quality expectation, then add relevant freight, duties, receiving, inspection, expected defects, and any other costs that differ between offers. Expected defect cost is the probability-weighted cost of the defect outcome, so its estimate deserves scrutiny. Include costs over the decision horizon, but avoid adding common costs that cannot change the choice. The cheaper total may still be unsuitable if capacity, delivery reliability, or specifications fail a required condition. When lead times or inventory ownership differ, include their financial effects if they are material. A management accounting guide can help reinforce the distinction between relevant incremental costs and costs that remain the same under either bid.

Worked example: For 1,000 acceptable units, supplier A quotes $10 each, with $1,000 freight and $500 expected rework: 1,000 × $10 + $1,000 + $500 = $11,500. Supplier B quotes $10.70 each, with $300 freight and $100 expected rework: $10,700 + $300 + $100 = $11,100. B costs $400 less on the stated basis, assuming both can deliver the required quantity and quality.

Watch the distinction: Choosing A from its lower $10 unit price ignores freight and expected rework. The comparison must use costs for the same quantity and service requirement.

Reference: [1]

30. Interpret a supplier scorecard

A supplier scorecard combines several performance dimensions, but it should retain the evidence behind the total. Define each measure, its observation period, and whether a higher score is better. Multiply each score by its weight, then sum the weighted results; weights must total 100% for a straightforward weighted average. Delivery, quality, responsiveness, cost, and risk can matter differently by material category. A single total may conceal a weak measure that breaches a minimum condition, such as a critical quality requirement. Examine trends and the number of observations before escalating: one late shipment and a persistent pattern need different responses. Use the scorecard to agree on corrective actions and verify improvement, rather than rewarding a high aggregate score while an essential service measure deteriorates.

Worked example: Suppose a supplier earns 90 for delivery, 80 for quality, and 70 for responsiveness, each on a 100-point scale. The respective weights are 50%, 30%, and 20%, totaling 100%. Its weighted score is 90 × 0.50 + 80 × 0.30 + 70 × 0.20 = 45 + 24 + 14 = 83 points. The total supports comparison, while responsiveness at 70 identifies a specific improvement need.

Watch the distinction: An 83-point total does not mean every service dimension is strong. A low score in an essential measure can require action despite the average.

Reference: [1]

31. Price a supply-risk response

A supply-risk response has a cost and an expected benefit, both of which depend on assumptions. Estimate the likelihood of a disruption and the loss if it occurs; probability multiplied by loss gives expected loss for a defined period. Compare that with the fees, premiums, qualification work, and residual disruption loss under a backup arrangement. The backup does not necessarily eliminate risk: two suppliers may depend on the same facility or transport lane. Expected values help compare repeated or portfolio decisions, while a single severe event may exceed the firm's risk tolerance even if its probability is low. Test the result against plausible probabilities and loss sizes, then check whether the alternate supplier can actually provide the needed quantity at the required time. The project risk guide provides related practice in assessing response costs and residual exposure.

Worked example: Assume a disruption has a 10% chance each year and would cause a $100,000 loss. Expected annual loss is 0.10 × $100,000 = $10,000. A qualified backup costs $6,000 annually plus an expected $2,000 price premium, totaling $8,000. If it eliminates that disruption loss, the stated expected-cost advantage is $2,000 a year. Recalculate if the backup shares the failure source or covers only part of demand.

Watch the distinction: A $10,000 expected loss is not a prediction that exactly $10,000 will be lost. Actual loss could be zero or much larger in one year.

Reference: [1]

32. Plan the inbound availability date

The date material can be used is later than the date a supplier finishes making it. Work backward from the required use date across supplier production, dispatch, transit, customs where relevant, receiving, inspection, and put-away. Sum stages that occur sequentially; do not add stages that truly overlap. State clearly whether dates mark the start or end of a week, because an ambiguous bucket can shift the release by a full week. Include any firm receipt already in transit rather than creating a duplicate order. When lead times vary, the deterministic offset gives a baseline but does not describe the chance of lateness; time buffers or alternate supply may be needed for critical material. The planning record should use the date inventory becomes usable, not simply the supplier's promised ship date.

Worked example: A purchase order is released at the start of week 2. Supplier production takes three full weeks, occupying weeks 2 through 4 and finishing at the start of week 5. Transit takes week 5, and receiving plus inspection takes week 6. The material is usable at the start of week 7. Planning it as available in week 5 would omit two required stages.

Watch the distinction: A supplier completion notice is not the same as usable inventory. Transit and receiving can delay material after production has ended.

Reference: [1]

33. Validate supplier capacity

A supplier's stated production capacity must be compared with its existing commitments over the same period and for the same resource. Ask whether the figure is demonstrated output or an ideal maximum, and account for maintenance, yield, labor, and shared equipment. Calculate residual capacity as available capacity minus committed load, then compare the new order with that remainder. If it does not fit, evaluate a staggered delivery, another qualified source, an approved overtime plan, or a change in customer timing. Capacity available in a later week cannot support an earlier due date without a feasible lead-time path. A supplier's promise is stronger when its material, labor, and bottleneck schedule substantiate it. Monitor actual output against the promised ramp to detect slippage before it becomes an inbound shortage.

Worked example: A supplier can produce 500 units in a particular week and already has 400 units committed on the same constrained line. Residual capacity is 500 - 400 = 100 units. A new 150-unit order would bring load to 550, exceeding capacity by 50 units. At least 50 units must move to another feasible period or source unless the supplier verifies 50 additional units of capacity.

Watch the distinction: The supplier's 500-unit weekly capacity is not free capacity. Existing 400-unit commitments leave room for only 100 more units in that week.

Reference: [1]

34. Control outsourced revisions

Engineering revision control links the approved product definition to procurement and supplier execution. Identify the effective revision by item, order, date or serial range, then ensure the purchase order, drawings, supplier acknowledgment, and inspection criteria agree. A supplier confirming an earlier revision creates a discrepancy even if price and delivery match. Determine whether work has started and whether old-revision material may be used under an explicit disposition; otherwise, correct the documents and supplier plan before production. Track any work in process, tooling, and inventory affected by the change, since a revision may alter components or testing. A blanket statement that the newest revision always applies can also be wrong when effectivity is tied to specific lots. The goal is traceable agreement on which specification governs each delivered unit.

Worked example: Suppose revision B is effective for a 500-unit order due next month, but the supplier acknowledges revision A. The confirmation conflicts with the applicable specification for all 500 units. Hold the discrepancy for resolution, send the correct revision and effectivity details, and obtain an updated confirmation before production. If any A units already exist, decide their disposition separately rather than counting them as compliant B supply.

Watch the distinction: A matching quantity and delivery date do not make a supplier confirmation correct. The revision determines what is being bought and inspected.

Reference: [1]

Plan and Manage Inventory

35. Distinguish inventory positions

Inventory records describe several distinct positions. On-hand quantity is physically present at a location; available quantity excludes stock already allocated, blocked, or otherwise unavailable. Inventory position for replenishment commonly adds open replenishment orders and subtracts backorders, so it can exceed on-hand stock. Raw material, work in process, and finished goods are also different forms of inventory: a finished unit cannot be counted again as its constituent raw parts. Track ownership separately when goods are consigned or held off-site. When checking an order promise, use stock in the correct item, revision, location, and usable status, plus receipts that arrive in time. When triggering a reorder point, use the position defined by that policy. Mixing these measures produces either duplicate supply or unnecessary orders.

Worked example: A storeroom physically holds 200 raw components. Fifty are reserved for released jobs, leaving 200 - 50 = 150 uncommitted raw components, assuming all others are usable. A separate finished-goods location holds 30 units available for customers. Those 30 are finished products, not additional raw components. If a new job needs 160 raw components, current uncommitted raw stock is short by 10.

Watch the distinction: Adding finished units to raw-component stock counts different items as interchangeable. A reservation also reduces uncommitted supply even though the components remain physically present.

Reference: [1]

36. Prioritize with annual usage value

ABC analysis ranks items by annual usage value: annual quantity consumed multiplied by unit value. Sort items from highest to lowest, calculate each item's share of total value, and inspect cumulative value to decide where tighter controls will pay off. High-value items often merit more frequent review, careful records, and closer supplier attention, while low-value items may use simpler replenishment controls. The boundaries between classes are management choices, not universal numerical rules. Value alone does not capture every operational risk: a cheap part that stops an entire line may still need strong protection. Consider criticality, lead time, theft risk, and demand variability alongside ABC ranking. Update the analysis when prices, usage, or product mix change, because yesterday's high-value item may become obsolete.

Worked example: Item X uses 1,000 units annually at $20 each, so its usage value is $20,000. Y uses 200 at $50, worth $10,000; Z uses 500 at $2, worth $1,000. Total annual usage value is $31,000. X represents $20,000 ÷ $31,000 ≈ 64.5% of value, Y about 32.3%, and Z about 3.2%. Review X first for value-focused controls.

Watch the distinction: ABC rank is not the same as unit price. Y costs more per unit than X, but X has twice Y's annual usage value.

Reference: [1]

37. Find the economic order quantity

Economic order quantity (EOQ) balances two costs that move in opposite directions under a deterministic model. Larger orders reduce the number of orders and annual ordering cost, but increase average cycle stock and annual holding cost. With annual demand D, cost per order S, and annual holding cost H per unit, EOQ is the square root of 2DS/H. Keep time units consistent: D and H must refer to the same annual period. EOQ assumes steady demand, a constant order cost and holding rate, and replenishment conditions that make average cycle stock about Q/2. It does not by itself decide when to order or how much safety stock to hold. Compare the calculated quantity with pack sizes, supplier minimums, capacity, shelf life, and quantity discounts before applying it. The CPIM practice questions can reinforce the calculation and its limits.

Worked example: Assume annual demand D = 10,000 units, ordering cost S = $50 per order, and annual holding cost H = $2 per unit. EOQ = √(2 × 10,000 × 50 ÷ 2) = √500,000 ≈ 707 units per order. At 707, annual ordering cost is about $707 and cycle-stock holding cost about $707. The result excludes safety stock, discounts, and storage constraints.

Watch the distinction: EOQ is an order quantity, not a reorder trigger. A separate timing rule must account for lead-time demand and any chosen safety stock.

Reference: [1]

38. Set a continuous-review reorder point

A continuous-review reorder point triggers replenishment when inventory position reaches the amount expected to be consumed during replenishment lead time plus a safety buffer. For stable average daily demand d and fixed lead time L, expected lead-time demand is d × L. Add the chosen safety stock to obtain the trigger. Inventory position generally equals usable on-hand stock plus open replenishment orders minus backorders; using on-hand alone can cause duplicate orders while goods are already inbound. The reorder quantity is a separate decision. If demand or lead time varies, the safety stock reflects the desired exposure to shortages and the observed variability. Review the policy when demand patterns, supplier performance, or service goals change, since an old point can become too low or tie up excess inventory.

Worked example: Assume average demand is 40 units per day, replenishment lead time is five days, and the chosen safety stock is 50 units. Expected lead-time demand is 40 × 5 = 200 units. Reorder point is 200 + 50 = 250 units of inventory position. When the position reaches 250, place the planned replenishment order; the 50-unit buffer covers demand beyond the stated average.

Watch the distinction: Ordering only when physical stock reaches zero leaves no stock for demand during the five-day lead time. The trigger must anticipate that demand.

Reference: [1]

39. Quantify a safety-stock choice

Safety stock protects against demand or replenishment uncertainty beyond the expected lead-time requirement. Under an approximately normal lead-time demand model, multiply the standard deviation of demand over the entire lead time by a selected one-sided service factor: safety stock = z × σLT. The standard deviation must already describe the lead-time period; using a daily deviation without conversion would understate the buffer. A higher factor reduces the chance of a stockout in a replenishment cycle but raises average inventory and carrying cost. This calculation targets a cycle-service probability under its assumptions; it does not directly specify unit fill rate. If demand is intermittent, lead time is highly variable, or the distribution is strongly skewed, test whether the normal approximation is suitable. Round to practical whole units and revisit the inputs as performance changes.

Worked example: Assume lead-time demand is approximately normal with a standard deviation of 30 units. Management selects a one-sided service factor of 1.65 for this hypothetical item. Safety stock is 1.65 × 30 = 49.5 units, rounded up to 50 whole units. Add this buffer to expected lead-time demand when setting the reorder point. The 50 units address variation, not average consumption.

Watch the distinction: Multiplying the factor by average demand would produce the wrong buffer. The factor applies to the standard deviation of demand over lead time.

Reference: [1]

40. Order under periodic review

A periodic-review system makes replenishment decisions at fixed review intervals. Because the next chance to order comes after the review interval, a new order must protect demand over the review interval plus replenishment lead time. With average daily demand d, review interval R, lead time L, and planned safety stock SS, an order-up-to target is d × (R + L) + SS. At each review, subtract inventory position from that target to obtain the order quantity, subject to minimums and pack sizes. Inventory position includes usable on-hand stock and open receipts, less backorders, so an order already on its way reduces the new order. The formula assumes the stated averages and timing are appropriate; variable demand or lead time calls for a buffer based on variation across the full protection period.

Worked example: Assume demand averages 20 units per day, reviews occur every seven days, lead time is three days, and planned safety stock is 40 units. The protection period is 7 + 3 = 10 days. Target inventory position is 20 × 10 + 40 = 240 units. If current inventory position, including open orders, is 100 units, order 240 - 100 = 140 units at this review.

Watch the distinction: Protecting only the three-day lead time overlooks demand before the next review. Periodic review must cover both intervals when setting its target.

Reference: [1]

41. Interpret service measures

Unit fill rate measures the share of requested units supplied immediately from available stock. Its denominator is units requested, so a large shortage can affect it sharply even if the shortage occurs only once. Cycle service level asks a different question: what proportion of replenishment cycles pass without any stockout? Each cycle counts once, regardless of how many units were missed. Identify the unit of measurement before comparing service figures or changing safety stock. A high cycle service level can coexist with a poor fill rate if one rare stockout is large; frequent small shortages can produce the opposite pattern. Both measures need a defined observation period and a consistent treatment of backorders. The CPIM content outline includes inventory service trade-offs and operational service measures.

Worked example: Suppose customers request 1,000 units during ten completed replenishment cycles, and 30 units cannot be filled immediately. Unit fill rate is (1,000 - 30) ÷ 1,000 = 97%. If all 30 missed units occurred in one cycle, nine cycles were stockout-free, giving 9 ÷ 10 = 90% cycle service. If shortages occurred in three cycles, cycle service would be 70%, while unit fill rate would remain 97%.

Watch the distinction: A 97% unit fill rate does not mean 97% of replenishment cycles avoided a stockout. The measures count different events and need separate underlying records.

Reference: [1]

42. Correct inventory records

A planning system can only net requirements correctly when its recorded balance matches usable stock. Cycle counting checks selected locations or items during normal operations, then compares the controlled physical count with the system quantity at a defined cutoff. Calculate variance as counted quantity minus recorded quantity. Before posting an adjustment, recount, check the unit of measure, and reconcile transactions around the cutoff; a receipt entered after the count can create a false discrepancy. Once a real variance is confirmed, correct the record and investigate its cause. Examine picking, put-away, scrap, transfers, labeling and unposted production issues. Repeated adjustments without finding the process fault leave future plans exposed. For lot-controlled stock, verify identity and status as well as quantity, since a correct total can still conceal unusable or misplaced material.

Worked example: At a transaction cutoff, a bin record shows 100 components. A controlled count finds 96, so variance is 96 - 100 = -4 components, or a 4% shortfall against the record. A recount confirms 96 and no pending transaction explains the difference. Adjust the record to 96, then review issue and movement records for the missing four; replenishment should use the corrected balance.

Watch the distinction: Do not change the record after a single unexplained count and stop there. A timing or unit error may be responsible, and an unresolved process fault will recur.

Reference: [1]

43. Disposition returned stock

A returned unit is an incoming item with uncertain condition, ownership and disposition, not automatically available supply. Record the receipt separately from saleable stock, retain its item and lot identity, and inspect against the relevant acceptance criteria. A cleared unit can return to saleable inventory; a repairable unit belongs in a controlled rework flow until it passes inspection; scrap has no usable supply value. Other dispositions may include refurbishment, recycling or return to a supplier, depending on condition and ownership. This separation protects order promising and material plans from counting stock that cannot satisfy demand. It also preserves traceability if a defect affects a particular lot. Compare recovery value with inspection, handling and repair costs when deciding whether a recovery path is worthwhile, while recording the physical quantities accurately at each step.

Worked example: A return shipment contains 20 units. Inspection clears 12 for resale, routes five to rework and scraps three; 12 + 5 + 3 = 20. Only 12 increase saleable available inventory now. The five can become available after successful rework and release. Crediting all 20 immediately would overstate current saleable supply by eight units and could cause false customer promises.

Watch the distinction: Receiving a return does not reverse the original sale into usable stock. Physical receipt and saleable release are separate transactions with different quality evidence.

Reference: [1]

Plan, Manage, and Execute Detailed Schedules

44. Build a routing-based job time

A routing describes the operations and resources a job must pass through. At each operation, setup time prepares the resource, while run time usually scales with quantity. Those times create work-center load. Queue time before processing, wait time between steps, and move time add to elapsed lead time but normally do not consume that work center's processing capacity. Keep the two clocks separate: capacity planning asks how many resource hours the job needs, while due-date planning asks when it can finish. If several operations are involved, add their resource loads by work center and follow their sequence to estimate elapsed time. Overlapping transfer batches or parallel operations can shorten elapsed time without removing all processing load. Check the routing's units and whether setup is per job, batch or unit before calculating.

Worked example: A 50-unit job uses one work center. Setup takes two hours and running takes 0.1 hour per unit, so run time is 50 × 0.1 = five hours and machine load is seven hours. Assume it also queues four hours and moves for one hour, with no overlap or other delays. Its elapsed routing time is 2 + 5 + 4 + 1 = 12 hours.

Watch the distinction: Treating all 12 elapsed hours as machine load would overstate capacity demand. Queue and move time delay completion, but they do not occupy this machine under the stated assumptions.

Reference: [1]

45. Load a finite schedule

Finite scheduling places work within the actual resource hours available in each period. Start with a calendar that reflects shifts, maintenance and staffing, then subtract already committed load. Rank ready jobs using due dates, material availability, priorities and setup effects; place each operation only where both the resource and required inputs are available. An infinite schedule can reveal an overload, but it does not resolve when the work will run. If a job can be paused or split, part may move into the next period; if processing must remain uninterrupted, schedule the entire operation in a later feasible slot or add capacity. Recheck downstream operations after any move because a locally feasible machine schedule can still miss the finished-order date. Record the decision so dispatching follows the feasible plan.

Worked example: A machine has eight available hours today. Two ready jobs need six and five hours, totaling 11, so today is overloaded by three hours. If the six-hour job runs first and the five-hour operation is safely divisible, two hours of the second job fit today and three hours carry forward. If it cannot be split, all five hours require another feasible slot.

Watch the distinction: An 11-hour load on an eight-hour day is not made feasible by listing both jobs on today's dispatch sheet. The overload needs a scheduling or capacity decision.

Reference: [1]

46. Compare sequencing rules

A sequencing rule decides which ready job uses a shared resource next. Shortest processing time tends to lower average completion and flow time when jobs are available together, processing is uninterrupted, and setup and due-date penalties do not dominate. Compute each job's completion time as the sum of all processing times before and including it; average those times across jobs. Earliest due date instead addresses deadline order, while setup-based grouping may save capacity at the expense of longer waits for some jobs. Choose the rule for the performance measure that matters, then test material readiness and committed dates. A rule that improves the average can still delay a critical customer order. For multiple work centers, local sequencing also needs a check against downstream constraints and work-in-process.

Worked example: Three jobs are ready at hour zero and need 2, 5 and 8 hours on one machine, with no setups or interruptions. Shortest-processing-time order completes them at hours 2, 7 and 15, averaging (2 + 7 + 15) ÷ 3 = 8 hours. Reverse order completes them at 8, 13 and 15, averaging 12 hours. The first sequence cuts average completion time by four hours.

Watch the distinction: Shortest processing time does not guarantee the best due-date performance. It minimizes this example's average completion time, but a long urgent job may need priority.

Reference: [1]

47. Protect the constraint

A constraint sets the maximum sustainable flow for the system when other resources have adequate capacity. Releasing work faster than it can process raises the queue without increasing completed output. The control sequence is to identify the constraint, keep it supplied with the right work, align upstream release with its pace, and improve or add constraint capacity when justified. A small time buffer before it protects throughput against ordinary upstream interruptions. Express that buffer in hours of constraint work, then convert to units using the constraint's rate. Monitor buffer penetration: persistent depletion signals starvation risk, while continued growth signals excess release or a changed constraint. Other resources may be idle at times without harming output. The aim is reliable system throughput, not maximum utilization at every individual work center.

Worked example: An upstream step can supply 80 units per hour, but the constraint processes 60 units per hour. With no other losses, releasing 80 creates 80 - 60 = 20 queued units each hour while output stays at 60. A 120-unit buffer equals 120 ÷ 60 = two hours of constraint work. Release toward the constraint's rate and replenish that buffer as it is consumed.

Watch the distinction: Keeping the faster upstream machine fully busy does not increase system output. Beyond the buffer needed to protect the constraint, its extra production becomes queue and longer lead time.

Reference: [1]

48. Dispatch by urgency

Critical ratio compares time remaining until a due date with work time remaining: time until due ÷ remaining processing time. A ratio above one indicates nominal slack; one means just enough time under the assumptions; below one signals that the remaining work exceeds the time left. Ratios help rank jobs on a dispatch list, but they are a warning signal rather than an automatic instruction to run the lowest number first. Check material availability, setup requirements, operation sequence and whether the remaining time includes queues or only direct processing. Recalculate as jobs move, because both the clock and remaining work change. When a ratio falls below one, examine recovery options such as an alternate routing, extra capacity, an agreed date change or a revised priority that protects the most important commitment.

Worked example: Assume both jobs are ready and their remaining processing times include all required operations. Job A is due in 12 hours and needs eight hours, so its critical ratio is 12 ÷ 8 = 1.5. Job B is due in six hours and needs eight, giving 6 ÷ 8 = 0.75. B has a nominal two-hour deficit and needs immediate recovery review.

Watch the distinction: A critical ratio below one does not mean the job is already finished late. It means the stated remaining processing time exceeds the time left before its due date.

Reference: [1]

49. Relate WIP to flow time

Little's law connects average work-in-process, throughput and flow time in a stable process: WIP = throughput rate × average flow time. Use consistent boundaries and units. Count the same jobs in WIP that appear in the completion rate, and use long-run averages rather than a brief snapshot during a surge. Rearranging gives flow time = WIP ÷ throughput. This relation is descriptive, not a promise that merely removing inventory will preserve output. If WIP falls below what a variable process needs to keep its constraint fed, throughput may also fall. Conversely, releasing work faster than the constraint completes it usually raises WIP and waiting time. Use the equation as a diagnostic, then inspect queues and release rules to find what is driving the change.

Worked example: A stable line averages 240 units in process and completes 80 units per day. Average flow time is 240 units ÷ 80 units per day = three days. If throughput remains 80 units per day while WIP rises to 320 units, implied flow time is 320 ÷ 80 = four days. The extra 80 units add one day of average time in the system.

Watch the distinction: Little's law does not prove that any WIP cut leaves throughput unchanged. Its comparison assumes a stable process and explicitly holds the completion rate at 80 units per day.

Reference: [1]

50. Size a pull signal

A kanban card authorizes a defined container of material in a pull system. Size the card count from expected demand during replenishment time plus a chosen buffer, divided by units per container; round up because a fraction of a card cannot authorize a full container. Replenishment time includes the relevant production, movement and waiting interval from signal to usable replacement. The chosen buffer covers stated variability or disruption risk, so changing lead time or demand should trigger a card-count review. Too few cards starve the consuming process; too many permit excess stock. Cards control authorized material in the loop, but they do not create capacity or guarantee a supplier response. For additional application practice, see CPIM practice questions.

Worked example: Assume demand of 100 units per day, a two-day replenishment time, a chosen 40-unit buffer and containers of 20 units. Demand during replenishment is 100 × 2 = 200 units. Add the buffer to get 240 units, then divide by 20 units per container: 12 cards. At most 12 containers, or 240 units, are authorized in this stated loop.

Watch the distinction: Using only one day's demand would understate the cards because replacement takes two days. The calculation must cover the full replenishment interval plus the chosen buffer.

Reference: [1]

Plan and Manage Distribution

51. Choose a distribution footprint

A distribution footprint determines where stock sits and how far orders travel. Compare network alternatives on a common annual basis, including facility expense, inbound and outbound freight, inventory carrying cost and handling where relevant. Then test whether each alternative can meet the promised delivery windows and available-space limits. More centers may shorten customer transit but split inventory across locations, potentially raising safety stock and transfer complexity. A single center can pool demand and reduce fixed expense, yet longer travel may weaken service or recovery from a disruption. Costs shared by all alternatives can be excluded from an incremental comparison if they truly stay equal. For uncertain events, make their assumptions visible and compare the resulting risk decisions with the required service, rather than treating the lowest freight figure as the whole answer.

Worked example: Assume both network options meet the same service requirement and all other annual costs are equal. One center incurs $120,000 in freight. Two centers reduce freight to $90,000 but add $40,000 in facility expense, totaling $130,000. The two-center option costs $130,000 - $120,000 = $10,000 more per year on the stated basis, so choose one center if no other benefit justifies that difference.

Watch the distinction: The $30,000 freight saving from a second center is not a $30,000 total saving. Its $40,000 additional facility expense more than offsets the freight reduction.

Reference: [1]

52. Time a distribution replenishment

Distribution requirements planning translates location-specific demand into time-phased replenishment. For each period, start with projected available stock, add confirmed inbound supply, and subtract requirements. If projected stock would fall below the chosen target, create a planned receipt using the stated lot-sizing rule. Move that receipt backward by the full replenishment lead time to find when the supplying location must release the shipment. This offset matters because stock arriving after the customer's required week does not prevent the shortage. Check whether on-hand quantity is actually uncommitted and whether scheduled receipts are reliable; otherwise the net requirement is understated. Repeat the calculation across periods because an earlier shipment can change later balances. The planned release also becomes demand at the supplying facility, where its inventory and capacity must be checked.

Worked example: A distribution center requires 100 units in week 4. Assume it has 30 uncommitted units on hand, a confirmed 20-unit receipt before week 4, no safety-stock target and lot-for-lot ordering. Net need is 100 - 30 - 20 = 50 units. Plan a 50-unit receipt for week 4. With one week of transport lead time, release the supplying shipment in week 3.

Watch the distinction: Releasing the 50 units in week 4 would deliver them one week late under the stated lead time. Netting identifies quantity; offsetting identifies the release date.

Reference: [1]

53. Compare transport time and cost

A transport choice must satisfy the delivery deadline before cost comparisons are meaningful. For each feasible mode, combine freight with relevant inventory carrying cost, handling, expected loss or damage and any stated service consequence. Transit time changes how long goods remain unavailable and may require more pipeline inventory. Calculate the cost of the time difference using the same shipment value and daily carrying rate; avoid charging the slower mode for a cost already included in both options. If the cheaper mode misses the customer window, it is infeasible unless the order can be rescheduled or the service consequence is explicitly valued and accepted. Reliability matters too: an advertised average transit time may be inadequate for a firm date if its variation is large. Compare modes using the actual need date and documented assumptions.

Worked example: Air freight costs $5,000 and takes two days. Ocean freight costs $1,000 and takes 12 days. Assume the ten additional transit days cost $100 per day in inventory carrying cost, with all other costs equal. Ocean's relevant cost is $1,000 + 10 × $100 = $2,000, $3,000 below air. If delivery is required by day 5, only air meets that deadline.

Watch the distinction: Choosing ocean from its $1,000 freight quote alone omits the extra $1,000 carrying cost. Choosing it for a day 5 requirement would also miss the stated deadline.

Reference: [1]

54. Measure order fulfillment accurately

Customer-service percentages are useful only when their denominator and success rule are clear. A line fill measure counts order lines supplied completely; a complete-order measure counts orders for which every line is supplied as required. One incomplete line can turn an entire multi-line order into an incomplete order, so complete-order performance may be lower even when most lines are filled. Neither measure is the same as unit fill rate, which counts requested units served immediately. Define whether the measure includes timeliness, quantity, product condition and the original promised date before interpreting it. Examine both the aggregate rate and the distribution of misses: a problem concentrated in one item or customer segment suggests a different fix from broad picking or transport failures. Keep the labels with the reported numbers so trends remain comparable.

Worked example: Assume 50 orders contain 100 lines in total. The warehouse supplies 90 lines completely, so line fill is 90 ÷ 100 = 90%. It supplies every line on 40 of the 50 orders, so complete-order performance is 40 ÷ 50 = 80%. The ten incomplete orders contain the ten incomplete lines in this example. Both percentages are correct but describe different outcomes.

Watch the distinction: A 90% line fill rate does not mean 90% of customers received complete orders. Multi-line orders fail the complete-order test if even one required line is missing.

Reference: [1]

55. Consolidate deliveries within service windows

Delivery consolidation combines stops or shipments to use vehicle capacity more efficiently. First define each customer's delivery window, load size and handling requirement. Build a route whose travel and service times fit those windows, then compare its total cost with separate deliveries. Savings can come from fewer dispatches and fuller vehicles, but an extra stop can make a later delivery infeasible or require inventory to wait. Check vehicle capacity, loading order, driver availability and any product incompatibility before accepting the route. If the combined route misses a promise, compare the cost of another departure or an approved new date rather than calling it an equivalent alternative. A route that saves money while keeping every agreed window is a valid operational improvement; one that breaks service needs an explicit trade-off decision.

Worked example: Three separate deliveries cost $100 each, so their total is 3 × $100 = $300. A proposed combined route costs $250 and, under the stated travel and unloading times, reaches all three customers within their promised windows. Consolidation saves $300 - $250 = $50. If any stop would fall outside its window, that $250 route would fail the stated service requirement.

Watch the distinction: A cheaper combined route is not automatically preferable. Its schedule must be checked stop by stop against promised windows and vehicle constraints before the $50 saving counts.

Reference: [1]

Manage Quality, Continuous Improvement, and Technology

56. Compare costs of quality

Quality costs fall into prevention, appraisal, internal failure and external failure. Prevention pays to avoid defects, such as process training or mistake-proofing. Appraisal pays to detect defects, such as inspection. Internal failure arises before delivery through scrap or rework; external failure arises after delivery through returns, claims or lost goodwill. A prevention proposal should be evaluated by its incremental cost against the failures and appraisal work it is expected to remove, using the same period and scope. Savings are estimates until observed, so distinguish a forecast from a measured result. A lower total cost can also improve delivery reliability by freeing capacity previously consumed by rework. If a proposal merely shifts detection later, apparent inspection savings may hide more costly customer failures. Cost analysis can help frame these comparisons.

Worked example: Current annual prevention cost is $5,000, appraisal is $3,000 and failures cost $12,000, for $20,000 total. Assume an added $2,000 prevention action leaves appraisal unchanged and reduces failures by $4,000. New cost is $7,000 + $3,000 + $8,000 = $18,000. The expected annual saving is $2,000, provided the projected failure reduction occurs.

Watch the distinction: An extra $2,000 spent on prevention is not automatically a cost increase overall. Compare it with the avoidable failure cost over the same period.

Reference: [1]

57. React to a process-control signal

A control chart compares a process statistic over time with limits estimated from process behavior. A point beyond a control limit is evidence of a possible special cause and calls for investigation of what changed around that observation. It does not, by itself, identify the cause. Customer specification limits serve another purpose: they define acceptable product or service results. A stable process can consistently produce some output outside specifications, while an unusual chart point can still fall within specifications. Choose a chart appropriate to the data, verify measurement and plotting, then examine equipment, materials, methods and staffing near the signal. Avoid changing a process after every ordinary fluctuation within control limits; such tampering can increase variation. After finding an assignable cause, remove or manage it and continue monitoring for renewed stability.

Worked example: Assume a valid control chart for a measured process characteristic has center line 10, lower control limit 8 and upper control limit 12. A new plotted statistic is 12.4, which is 0.4 above the upper limit. Flag it for a special-cause investigation and verify the measurement. No specification limits are given, so this chart alone cannot establish whether the resulting product is acceptable.

Watch the distinction: A point outside a control limit is not automatically a rejected product. Control limits describe process behavior; acceptance requires separately stated specification limits and an appropriate product measurement.

Reference: [1]

58. Interpret process capability

Process capability compares a stable process's spread and centering with specification limits. Cp divides specification width by six process standard deviations; it describes potential fit if the process is centered. Cpk uses the smaller distance from the process mean to either specification limit, divided by three standard deviations, so it reflects the nearer side. Cpk cannot exceed Cp for the same inputs. Before using either index, check process stability, a meaningful measurement system and whether the distributional assumptions behind the simple formulas are suitable. An unstable process makes a single mean and standard deviation poor predictors of future output. These indices summarize a relationship; they do not replace actual defect evidence or customer acceptance rules. A gap between Cp and Cpk is a cue to investigate process centering as well as variation.

Worked example: Assume a stable process with lower specification 9, upper specification 11, mean 10.2 and standard deviation 0.2 in the same measurement units. Cp = (11 - 9) ÷ (6 × 0.2) = 2 ÷ 1.2 = 1.67. Cpk is the smaller of (11 - 10.2) ÷ 0.6 = 1.33 and (10.2 - 9) ÷ 0.6 = 2.00. Thus Cpk = 1.33; its lower value shows the mean sits nearer the upper limit.

Watch the distinction: Cp of 1.67 does not show that the process is centered. Cpk uses the nearer specification limit and exposes the loss of margin caused by the 10.2 mean.

Reference: [1]

59. Verify an improvement cycle

Plan-do-check-act turns a proposed fix into a controlled learning cycle. In plan, define the defect, likely cause, intervention and measure, including a baseline with a clear denominator. In do, trial the change under documented conditions. In check, compare rates rather than raw counts when production volumes differ, and look for other changes that could explain the result. In act, standardize a credible improvement or revise the hypothesis and test again. A before-and-after difference is evidence, but one trial may reflect product mix, material lot or random variation. Keep the same defect definition and inspection method across periods. Continue monitoring after rollout, because a gain can fade when staffing, equipment or demand changes. The decision is strongest when the process mechanism and measured change point in the same direction.

Worked example: A setup-method trial compares two runs of 1,000 units each, inspected with the same defect definition. The earlier run has 40 defects, or 40 ÷ 1,000 = 4%. The trial has 15, or 1.5%. The absolute drop is 2.5 percentage points; the relative reduction is (4 - 1.5) ÷ 4 = 62.5%. If the runs are comparable, standardize the method and monitor later results.

Watch the distinction: A 62.5% observed reduction is not a guarantee for every later run. Check comparable conditions and continue measuring after the new method becomes standard work.

Reference: [1]

60. Trust the transaction before the plan

Enterprise resource planning records transactions and master data that planning systems use as inputs. An advanced planning system can optimize dates and quantities only against the balances, lead times and constraints it receives. A manufacturing execution system records shop-floor events and progress, but it cannot make nonexistent material available. Trace a suspect planning result back through the source transaction, item and location before changing replenishment parameters. Reconcile physical stock, receipts, issues, reservations and interfaces at a common cutoff; then correct the erroneous transaction with an auditable entry and repair its cause. Rerun the affected plan and inspect downstream orders or promises. The same logic applies to duplicated receipts, missing scrap postings and wrong units of measure. Better algorithms do not compensate for inaccurate inventory records, and automation can spread a bad record faster.

Worked example: Assume physical stock is 100 units, but a duplicate 10-unit receipt makes ERP show 110. A lot-for-lot planning run needs 150 units and assumes no other supply or safety stock. Using ERP data, it plans 150 - 110 = 40 units; using true stock, it needs 150 - 100 = 50. The error understates replenishment by ten units. Reverse the duplicate transaction, fix its interface cause and rerun the plan.

Watch the distinction: Do not treat the planning system's 40-unit order as independent evidence of available stock. It inherited the duplicate receipt and will repeat the shortage until the source record is corrected.

Reference: [1]

Official references and further reading

ascm-cpim9-current-ecm

All study guides

Align the Supply Chain to Support the Business Strategy

1. Read the competitive setting

2. Qualifiers, winners and customer value

3. Choose the push-pull boundary

4. Make or buy with the constraint in view

5. Match process and layout to volume

6. Measure strategy and sustainability

7. Map and mitigate supply risk

Conduct Sales and Operations Planning (S&OP) to Support Strategy

8. Place S&OP in the planning hierarchy

9. Build the portfolio demand view

10. Balance aggregate supply and inventory

11. Size a strategic capacity buffer

12. Reconcile the financial plan

13. Approve a feasible volume and mix

Plan and Manage Demand

14. Separate independent and dependent demand

15. Adjust a forecast for a market event

16. Update with exponential smoothing

17. Apply a seasonal demand pattern

18. Detect error and bias

19. Consume the forecast with orders

20. Shape demand within capacity

Plan and Manage Internal Supply Sources

21. Project the master schedule

22. Promise only available supply

23. Check rough-cut capacity

24. Explode the bill of materials

25. Net and offset MRP requirements

26. Calculate work-center load

27. Schedule final configuration

Plan and Manage External Supply Sources

28. Choose a sourcing structure

29. Compare total acquisition cost

30. Interpret a supplier scorecard

31. Price a supply-risk response

32. Plan the inbound availability date

33. Validate supplier capacity

34. Control outsourced revisions

Plan and Manage Inventory

35. Distinguish inventory positions

36. Prioritize with annual usage value

37. Find the economic order quantity

38. Set a continuous-review reorder point

39. Quantify a safety-stock choice

40. Order under periodic review

41. Interpret service measures

42. Correct inventory records

43. Disposition returned stock

Plan, Manage, and Execute Detailed Schedules

44. Build a routing-based job time

45. Load a finite schedule

46. Compare sequencing rules

47. Protect the constraint

48. Dispatch by urgency

49. Relate WIP to flow time

50. Size a pull signal

Plan and Manage Distribution

51. Choose a distribution footprint

52. Time a distribution replenishment

53. Compare transport time and cost

54. Measure order fulfillment accurately

55. Consolidate deliveries within service windows

Manage Quality, Continuous Improvement, and Technology

56. Compare costs of quality

57. React to a process-control signal

58. Interpret process capability

59. Verify an improvement cycle

60. Trust the transaction before the plan

Official references and further reading

Frequently Asked Questions

Related Study Guides

Certified Investment Management Analyst (CIMA - Investments)

Certified Management Accountant (CMA)

Wealth Management Certified Professional (WMCP)