How do a demand forecast and a supply plan differ?

A forecast estimates what customers may request. A supply plan decides how the organization will respond, given inventory, capacity, suppliers, lead times and financial limits. Sales and operations planning reconciles the two at an aggregate level. If forecast demand is 1,200 units but feasible output is 1,050, the gap is 150 units; planners must change capacity, inventory, timing or the demand plan rather than treat the forecast as an executable schedule.

Why can the lowest quoted supplier price be the more expensive choice?

The quote omits costs such as freight, duties, inspection, defects, inventory and disruption. Compare suppliers on the same delivered quantity and service requirement. For a hypothetical order of 1,000 usable units, a $9 quote with $1,500 in freight and expected quality costs totals $10,500. A $10 quote with $300 in those costs totals $10,300, making the higher quoted price the lower total cost under these assumptions.

How should service and inventory be balanced?

Higher inventory can buffer variation, but it ties up cash and may become obsolete. Separate cycle stock, which supports routine replenishment, from safety stock, which covers uncertainty. Then compare the cost of additional stock with the consequences of a shortage for each product or customer segment. A critical spare with unpredictable demand may justify a different buffer from a short-lived promotional item, even when their average demand is identical.

CSCP Study Guide: Decisions Across the Supply Chain

Start with concept 1 Choose a topic

A supply chain decision rarely ends where it starts. A promotion changes demand, which changes material requirements, capacity, transport and cash needs. A cheaper supplier can raise total cost if deliveries are unreliable or defects disrupt production. This guide follows those connections across the eight domains of the APICS Certified Supply Chain Professional (CSCP) scope. Each planned lesson centers on a decision you could face as a supply chain professional: what information to gather, which calculation or framework to use, what the result means and what constraint might change the answer. The examples use stated assumptions so you can practice reasoning from facts rather than memorizing a single preferred policy.

Forecast and Manage Demand

1. Read demand history carefully

Sales history contains several kinds of signal. A level is the recurring volume around which demand moves; a trend is sustained growth or decline; seasonality is a pattern tied to a recurring calendar period. A promotion, stockout or unusual customer order can distort all three. Start by marking such events on the time series, then compare like periods and ask whether recorded sales represent demand: sales during a stockout understate what customers wanted. Adjust exceptional observations only when the event and its effect are documented. Keep the original figures so the adjustment can be challenged later. The CSCP content outline includes historical analysis and demand patterns. The practical decision is whether a pattern is likely to recur before using it to set future supply.

Worked example: Monthly sales are 100, 120 and 140 units. The third month includes a documented clearance order of 30 units, so its adjusted recurring sales are 140 − 30 = 110 units. The adjusted sequence, 100, 120, 110, no longer supports a steady 20-unit monthly rise or a projection of 160. Investigate the second month's increase before choosing a baseline.

Watch the distinction: Treating every recorded sale as repeatable demand creates false trends. An exceptional order may be real revenue while still being unsuitable for a recurring forecast.

Reference: [1]

2. Scan demand outside the business

An internal order history describes what customers bought under past conditions. An environmental scan asks whether those conditions are changing. Relevant signals include competitor capacity, customer plans, channel inventory, economic conditions and changes in product availability. Translate a signal into a forecast adjustment only after specifying its causal path, likely timing and overlap with other signals. A competitor's closure matters only if its customers can buy your product through an accessible channel. A partner's inventory reduction might mean weaker consumer demand or merely a temporary order correction. Build a base case and explicit alternatives, then assign someone to validate each assumption. This keeps the forecast useful for capacity decisions while exposing uncertainty. Do not add every optimistic market estimate to the baseline as though the events were confirmed and independent.

Worked example: A retailer's usual run rate suggests 500 units next month. A competitor closure could redirect 80 units, while a channel partner expects a 20-unit decline. The scenario estimate is 500 + 80 − 20 = 560 units. Sales must check whether the redirected customers can reach the retailer and whether the partner decline overlaps with the competitor effect before committing supply for all 560.

Watch the distinction: External signals are not automatically incremental demand. Adding two reports that describe the same customer shift counts that demand twice and overstates the required supply.

Reference: [1]

3. Shape demand across the product life cycle

Demand shaping changes when, where or whether customers buy through the four Ps: product, price, place and promotion. The same action has different effects across a product's life cycle. At launch, limited awareness and uncertain adoption make comparable products and early customer evidence valuable. At maturity, established volume helps measure promotion lift, but a discount can move purchases between weeks without increasing total consumption. In decline, apparent price sensitivity may coexist with shrinking distribution or replacement by another product. Separate truly incremental units from purchases pulled forward, shifted from another channel or taken from another product in the portfolio. Plan replenishment for the timing effect as well as the total. Check margins and stock availability before calling a promotion successful: more units sold can still reduce contribution or create later excess stock.

Worked example: A mature item normally sells 200 units each week. A promotion is expected to pull 60 purchases from the following week into this week. Plan 200 + 60 = 260 units now and 200 − 60 = 140 next week. The two-week total stays 400 units. The first week's 30% increase therefore does not demonstrate lasting market growth.

Watch the distinction: Projecting a promotional spike into every later period mistakes timing for growth. Check the following period and total contribution before raising the permanent forecast.

Reference: [1]

4. Choose a forecasting method

Intrinsic forecasting uses patterns within the demand series, such as moving averages, smoothing, trend and seasonality. It works best when the underlying process is reasonably stable and enough comparable history exists. Extrinsic forecasting uses explanatory variables outside that series, such as customer contracts, market activity or a known price change. Qualitative judgment is especially useful for launches, disruptions and structural changes with little relevant history, but its assumptions should be recorded. Choose a method by forecast horizon, data quality, demand pattern and the decision it supports. A short, stable replenishment decision may need a simple average; a capacity investment may need scenarios tied to market drivers. Compare methods against held-out periods where possible, then adjust for verified future events that historical data cannot contain.

Worked example: Sales in three comparable periods are 90, 100 and 110 units. A three-period moving average forecasts (90 + 100 + 110) ÷ 3 = 100 units. A firm contract adds 40 units in the next period, making an adjusted forecast of 140. Confirm that those contract units are additional to normal sales; otherwise the adjustment double counts demand.

Watch the distinction: A mathematically tidy time-series result can be less useful than an adjusted forecast when a confirmed future event changes demand. The adjustment still needs evidence.

Reference: [1]

5. Measure error and bias

Define the error sign before calculating it. With error equal to actual demand minus forecast, positive errors mean underforecasting and negative errors mean overforecasting. Mean absolute error averages error magnitudes, so opposite signs cannot cancel; cumulative error preserves signs and helps reveal persistent bias. For forecasts across products or periods, compare like units and beware percentage errors when actual demand is near zero. Track errors by horizon because a one-week forecast and a six-month forecast serve different decisions. Investigate a run of same-sign errors for a systematic cause, such as omitted promotions, changed lead times or a stale baseline. Random error calls for uncertainty buffers or a better model; persistent bias calls for correcting the forecast process. For practice applying these distinctions, use CSCP practice questions.

Worked example: Forecasts are 100 and 120 units; actual demand is 110 and 130. Using actual minus forecast gives errors of +10 and +10 units. Mean absolute error is (|10| + |10|) ÷ 2 = 10 units, and cumulative error is +20 units. Both periods were underforecast, so review the baseline or any omitted demand driver.

Watch the distinction: A low average signed error does not prove accuracy: positive and negative errors can cancel. Examine absolute error for size and cumulative error for direction.

Reference: [1]

6. Reconcile demand and supply in S&OP

Sales and operations planning reconciles an aggregate demand view with inventory, production, purchasing, capacity and financial limits over a common planning horizon. First establish one demand plan, then calculate available supply from beginning inventory and feasible output. Quantify each gap by period rather than relying on a total that may hide a late delivery. Test options such as overtime, subcontracting, inventory use, demand timing or a revised service commitment. Each option needs a cost, a capacity and material check, and an accountable owner. Finance tests whether the proposed response supports the business plan; operations confirms execution; sales owns the customer implications. Record the agreed plan and assumptions, then review deviations at the next cycle. A nominally balanced plan is not feasible if its required components or labor are unavailable.

Worked example: Next month's demand is 1,000 units. Beginning finished goods are 100 and regular production can deliver 800, leaving 1,000 − 100 − 800 = 100 units uncovered. If approved overtime can produce 100 units and materials are available, total supply reaches 1,000. Operations owns overtime feasibility, purchasing confirms inputs, and finance adds its stated cost before approval.

Watch the distinction: Closing an aggregate arithmetic gap does not establish an executable plan. Output scheduled after the required date, or without materials, cannot satisfy that demand.

Reference: [1]

Manage the Global Supply Chain Network and Information

7. Design for the required service

Network design begins with the service promise and the demand locations it must cover. Candidate facilities and transport lanes are then tested for response time, capacity, fixed and variable cost, and disruption exposure. A low-cost site can be ruled out by a binding delivery requirement; among feasible choices, compare total cost and resilience. More locations can shorten the last mile but add rent, labor, stock duplication and coordination. Demand density and transport reliability matter as much as a warehouse's distance on a map. Specify whether a service target applies to orders, units, customers or geographic areas before measuring coverage. Document these requirements clearly; techniques from business analysis help distinguish a measurable service need from a vague request for faster delivery.

Worked example: A single warehouse costs $200,000 per year and can serve 80% of orders within two days. Adding a second costs $70,000 more and raises coverage to 96%. If the requirement is at least 95% of orders within two days, the single-site network is infeasible. The two-site option costs $270,000 yearly and clears the stated coverage test, pending capacity and resilience checks.

Watch the distinction: Selecting the cheapest network before checking service constraints can produce an unusable design. Compare costs only among alternatives that meet the required promise.

Reference: [1]

8. Map product, information and funds

A supply chain moves physical product, information and funds, often in different directions and at different times. An order communicates intent; shipment and receipt events establish what moved; an invoice and payment settle value. The records must refer to the same item, quantity, unit of measure and transaction. A mismatch can make available inventory look wrong, trigger an unnecessary replenishment, or release payment for goods not received. Trace the discrepancy through shipping evidence, carrier records, receiving count and system postings before correcting either the physical or financial record. Funds may legitimately move under agreed payment terms, but the business still needs a controlled exception process. Mapping these linked flows reveals whether a delay starts with goods, a missing event message or a billing dispute.

Worked example: A supplier reports shipping 500 units, while the receiving message records 450. The unresolved difference is 500 − 450 = 50 units. Check the physical count, shipment documents and receipt posting before approving payment for all 500. If 500 arrived, correct the receipt record; if only 450 arrived, investigate the 50-unit shortfall and settle the invoice accordingly.

Watch the distinction: A shipment notice is evidence of dispatch, not proof of receipt. Paying or replenishing from that single event can compound an unresolved quantity error.

Reference: [1]

9. Build usable end-to-end visibility

Visibility is useful when an event can be linked to an item, location, order and decision deadline. Shipment status alone tells a planner little unless it identifies which commitments are at risk. Establish shared identifiers and timely event messages for milestones such as dispatch, border clearance, arrival and receipt. When an exception occurs, compare the new arrival estimate with promised dates, available stock and alternative supply. Then prioritize according to agreed service rules and communicate what cannot be covered. Data quality matters more than the number of dashboard fields: a late but accurate estimate may still support action, while an unverified location ping can mislead. Define who receives alerts and who may change customer commitments so the information produces a coordinated response.

Worked example: A carrier reports that 200 units will arrive two days late. Linked order records show 120 urgent units can be supplied from available stock. Allocate those 120 after confirming stock is uncommitted, leaving 200 − 120 = 80 units exposed. Contact the affected customers about the 80 units and check whether an alternate lane or substitute stock can close the remaining gap.

Watch the distinction: Collecting more tracking events does not itself improve service. The event must reach a decision maker with enough order and inventory context to act.

Reference: [1]

10. Protect connected supply chain systems

Integration lets partners exchange orders, shipment updates and inventory information quickly, but it also extends the consequences of bad data or a compromised account. Grant access by business role and the minimum actions required: reading status, posting a shipment event and changing payment details are different privileges. Authenticate users, review permissions when roles change, and log important edits so records can be traced. Validate inbound data for valid item identifiers, quantities and formats before it drives planning or payment. Design recovery around the transactions the business must restore first, with tested backups and a controlled manual process for urgent work. A connected system is only as dependable as the identities, master data and recovery procedures around its interfaces. Evaluate a partner connection against the specific information flow it needs.

Worked example: A carrier partner needs to see shipment status but has no role in supplier bank details. Give its account read-only access to relevant shipment records and no payment-data permissions. If that account is compromised, the attacker may view permitted status data but cannot use that access to alter bank instructions. Log access and review the scope when the arrangement changes.

Watch the distinction: Giving every integration broad edit rights for convenience turns one account failure into a wider financial and data-integrity risk. Permissions should match actual tasks.

Reference: [1]

11. Govern master data

Master data supplies the common definitions behind transactions: item codes, units of measure, customer addresses, supplier identities, lead times and approved specifications. Create a record with an owner and required fields, validate it before use, and control later changes. Cleansing finds duplicates, obsolete entries and conflicting values; retirement prevents old records from appearing in new orders while preserving transaction history. Unit conversions deserve special attention because planning, purchasing and warehousing may use different units for the same item. An incorrect conversion can make stock appear sufficient when it is not, or inflate an order by a full case multiple. Track the effective date of changes and communicate them to connected systems. Governance is successful when every process interprets the same item and partner in the same way.

Worked example: An item is purchased in cases of 12 and stocked in individual units. A purchase order for 50 cases represents 50 × 12 = 600 individual units. If the receipt interface posts only 50 units, the stock record understates the receipt by 550 units. Correct the conversion and reconcile the physical receipt before using that balance to plan orders.

Watch the distinction: A consistent item code does not guarantee consistent quantities. A case-to-each conversion error can survive across systems unless units are checked at each handoff.

Reference: [1]

12. Design a balanced performance view

A balanced performance view links customer outcomes to the resources consumed to achieve them. Operational measures include delivery reliability, responsiveness and quality; financial or asset measures include cost, inventory investment and working capital. The Supply Chain Operations Reference model offers a common structure for comparing such dimensions across connected processes. Define each metric's numerator, denominator, time period and data owner before putting it on a dashboard. A service gain may be valuable, but it can conceal a large stock increase or expensive expedites. Compare changes against the business objective and the marginal cost of improvement, ideally by customer segment. Look for cause and effect rather than optimizing one number independently: raising inventory at every site may improve availability while weakening asset efficiency and increasing obsolescence.

Worked example: On-time delivery rises from 92% to 97%, a five-percentage-point gain. Average inventory rises from $1.0 million to $1.4 million, an increase of $400,000 or 40%. Report both changes. Management can then estimate the value of the service gain and the carrying and obsolescence costs of the extra stock before deciding whether this is an improvement.

Watch the distinction: A single improving service metric does not prove the network improved overall. Pair it with the cost and asset measures that changed to produce that service.

Reference: [1]

Source Products and Services

13. Make or buy by total cost and capability

A make-or-buy decision compares relevant future costs and capabilities for a specified volume and time horizon. Internal cost may include avoidable labor, materials, setup and any capacity that could be used elsewhere; external cost may include price, qualification, freight, quality and supplier management. Do not allocate unavoidable overhead as though buying would eliminate it. Test both choices against required quality, lead time, intellectual property protection and core competencies. A cheaper supplier is not a viable alternative if it cannot meet the specification or threatens a strategically important capability. Capacity also changes the answer: buying can relieve a constrained plant even when the purchase price exceeds unit manufacturing cost. State the volume assumption and recalculate if demand changes, since fixed qualification or tooling costs spread differently across volumes.

Worked example: For 10,000 units, internal production costs $40,000 in relevant fixed cost plus $6 × 10,000 = $60,000 variable cost, totaling $100,000. Buying costs $8 × 10,000 = $80,000 plus $10,000 qualification, totaling $90,000. Buying saves $10,000 on these assumptions if the supplier passes capability, quality and strategic checks.

Watch the distinction: Comparing only the $6 make cost with the $8 buy price ignores the stated fixed cost. Equally, including unavoidable overhead would overstate savings from outsourcing.

Reference: [1]

14. Specify sourcing needs and timing

Sourcing requirements combine what is needed with when it must be usable. Define the specification, unit of measure, quantity, quality evidence, delivery location and required receipt date. Then work backward through every lead-time step, including supplier qualification, approvals, production, inspection and transit. Some steps can overlap, but only when the process actually permits it. The release date is the latest date at which the first required step must begin, subject to a chosen contingency allowance. Compare the calculated date with supplier capacity and internal approval timing; a nominal production lead time alone often hides the earliest commitment needed. If the date has already passed, escalate a real option such as an alternate qualified source, changed demand timing or expedited transport rather than silently entering an impossible order date.

Worked example: A component must be available at the start of week 10. Supplier qualification takes two weeks, production three and transit one, with no overlap assumed. Total lead time is 2 + 3 + 1 = six weeks, so qualification must start at the start of week 4. Starting in week 7 would leave only three weeks and cannot meet week 10 under these assumptions.

Watch the distinction: Subtracting only transit or production time from the need date omits upstream work. An order can appear timely in the system while qualification makes delivery impossible.

Reference: [1]

15. Segment purchasing categories

Category segmentation uses at least two questions: how much the purchase affects business results, and how difficult supply is to replace. Impact may reflect spend, production dependency or customer consequences; supply risk may reflect scarce capacity, unique specifications or switching time. Routine, low-risk goods suit simple catalogs and efficient transactions. High-impact goods with competitive supply offer opportunities to compare bids and negotiate total value. Items that are hard to source need continuity measures, even when annual spend is modest. Strategic items combine high impact and high risk, calling for close planning, supplier development and explicit risk sharing. Revisit the classification when demand, design or markets change. The point is to direct management effort where it changes outcomes, rather than give every purchase the same approval path.

Worked example: A common fastener costs little, has several qualified sellers and can be replenished quickly, so a catalog with standard terms controls transaction effort. A proprietary chip has lower annual spend but stops a major production line if unavailable. Its limited alternatives and high consequence justify capacity reviews, approved substitutes or redesign work with the supplier.

Watch the distinction: Annual spend alone misses criticality. A low-spend component can deserve intensive management when a shortage halts production and replacement takes a long time.

Reference: [1]

16. Analyze the supply base

Supply base analysis combines spend records with capacity, performance, location and dependency data. Normalize supplier names and item definitions first, since duplicates can hide concentration or fragmented buying. Consolidating volume can improve terms and reduce administrative work, but the proposed supplier must have verified capacity and a credible ramp-up plan. Examine what happens if that supplier fails, including whether another qualified source can respond within the required time. Right-sizing can mean adding a source for resilience as well as removing unnecessary ones. Compare total cost, quality and lead time after consolidation, not just quoted unit prices. For a constrained supplier, reserve capacity by item and period so the plan does not assume the same production slot serves multiple customers.

Worked example: Three suppliers each deliver 100 units per week, for 300 total. The lowest-price supplier has verified capacity of 180 units weekly. Moving all demand there exceeds capacity by 300 − 180 = 120 units per week. Keep at least 120 units with other qualified suppliers unless capacity is expanded and demonstrated before transfer.

Watch the distinction: A price ranking is not a capacity plan. Awarding all volume to the lowest bidder can create an immediate shortage if its available output is below demand.

Reference: [1]

17. Influence design before sourcing

Early procurement input can change a product's lifetime supply cost while design choices remain flexible. Ask whether standard materials, common parts or easier manufacturing can meet the required performance. Packaging design affects pallet use, warehouse handling, damage risk and return logistics; it is therefore part of total cost, not merely a purchase-price detail. Compare alternatives using validated specifications and downstream costs. A denser package may reduce freight but fail protection tests, increase handling time or make recycling harder. Work with engineering, operations, logistics and suppliers to test these effects before approval. Document why a custom feature is necessary and whether a standard alternative preserves safety, quality and customer experience. This is a design decision with sourcing consequences, not an instruction to substitute cheaper materials without qualification.

Worked example: A standard package holds 100 units per pallet; a custom one holds 80. Shipping 1,000 units needs 1,000 ÷ 100 = 10 standard pallets or 1,000 ÷ 80 = 12.5, rounded up to 13 custom pallets. The standard design saves three pallets on this shipment if protection tests pass and handling requirements remain acceptable.

Watch the distinction: Packing more units onto a pallet is not automatically better. Damage, safety and handling effects can cost more than the three-pallet saving.

Reference: [1]

18. Qualify a supplier

Supplier qualification tests whether a bidder can repeatedly deliver the required product under normal and stressed conditions. Begin with the specification and critical characteristics, then examine process capability, measurement controls, quality history, capacity and management systems. Financial health, ownership and compliance evidence matter when failure would interrupt supply. A sample that passes inspection is useful but does not establish stable production: review how variation is controlled and how nonconforming output is contained. For critical tolerances, verify that equipment, methods and operators can measure and hold the requirement. Record gaps and require evidence of correction before committing essential volume. Once suppliers meet mandatory requirements, compare total cost and service. This sequence prevents an attractive bid from winning on price while transferring unmeasured defect or continuity risk to the buyer.

Worked example: Supplier A quotes $2 less per unit than Supplier B but cannot show a validated process for a critical tolerance. Supplier B demonstrates controlled measurements and production records meeting the specification. B is the currently qualified choice. A's potential saving should be reconsidered only after it proves capability and the buyer calculates the full cost of either option.

Watch the distinction: A low bid or one clean sample does not prove sustained capability. Qualification asks whether the supplier's process can keep meeting the specification at planned volume.

Reference: [1]

19. Compare delivered sourcing cost

Delivered sourcing cost starts with purchase price and adds costs caused by obtaining usable goods. Depending on the decision, include freight, duties, inspection, expected defects, rework, carrying cost from lead time and inventory, and disposal. Use the same demand volume and cost boundary for each supplier. Expected defect cost is probability multiplied by the cost of each defect, or an estimate from comparable performance; it is uncertain, so test sensitivity. A supplier with a higher unit price can be cheaper overall if transport and quality costs are lower. Also check whether a quoted cost is one-time or recurring. The management accounting guide can help reinforce the distinction between relevant incremental costs and allocated costs when evaluating alternatives.

Worked example: For 1,000 units, A costs 1,000 × $10 = $10,000 in goods, plus $800 freight and $400 expected defects, totaling $11,200. B costs 1,000 × $10.50 = $10,500, plus $300 freight and $100 expected defects, totaling $10,900. B is $300 cheaper on the stated costs, despite its higher unit price. Add any material carrying-cost difference before final selection.

Watch the distinction: Choosing the lowest unit price ignores costs incurred after the purchase order. Compare suppliers at the same usable volume and state which cost elements are included.

Reference: [1]

20. Negotiate workable contract terms

A workable contract allocates responsibilities that each party can control and makes operational expectations measurable. Price belongs alongside quantity commitments, delivery terms, quality acceptance, change procedures, remedies and any value-added services. A discount tied to minimum orders may shift inventory cost and obsolescence risk to the buyer. Conversely, a flexible release schedule may cost more per unit but better match uncertain demand. Define how forecasts differ from binding orders and what happens when requirements change. Delivery terms should identify the point of responsibility for freight and relevant risks so landed cost can be compared consistently. Evaluate a proposed concession against its whole economic effect and the likelihood it can be executed. Escalation and correction provisions should support recovery when performance slips, not merely assign blame afterward.

Worked example: Annual purchases at current terms total $100,000. A 3% discount saves 0.03 × $100,000 = $3,000, reducing the price component to $97,000. The required minimum orders add $5,000 in annual holding cost. Stated total cost becomes $102,000, which is $2,000 above the original $100,000. Negotiate smaller releases or reject the concession on these assumptions.

Watch the distinction: A discount is not necessarily a saving. If its quantity terms create more carrying cost than the price reduction, the buyer pays more overall.

Reference: [1]

21. Choose a purchase-order approach

Use a standard purchase order when quantity, specification, price and delivery date are defined for a particular purchase. A blanket arrangement establishes terms for repeated purchases, with releases specifying what to deliver and when. It reduces repetitive ordering work, but the agreement may still impose minimum quantities, notice periods or firm delivery windows. Before changing a release, distinguish the forecast from a binding commitment and check whether materials have already been made. Track acknowledgments and receipts against each release so an expedite addresses a genuine shortage, while a de-expedite prevents unwanted stock. Electronic purchasing can automate routine steps when item, supplier and approval data are dependable. The CSCP content outline places standard and blanket orders, order changes and expedites within purchasing management.

Worked example: A blanket agreement covers 1,200 units for a year, initially released at 100 units monthly. Six months have passed, so 600 units remain on the original schedule. Demand falls to 80 monthly for the six remaining months: 6 × 80 = 480 units. Revised releases would reduce expected receipts by 120 units, if the agreement permits that change.

Watch the distinction: A blanket agreement is not permission to change every release freely. Minimum purchases, firm periods and supplier commitments can make a reduced schedule costly even when demand has fallen.

Reference: [1]

22. Reconcile the purchase-to-pay record

Purchase-to-pay control compares three different claims: the purchase order states what was authorized, the receipt records what arrived and the invoice requests payment. Match item, unit of measure, quantity, price and applicable charges before treating an invoice as payable. A mismatch may reflect a short delivery, an unrecorded receipt, a price change or a data-entry error; each calls for different evidence. Quarantined or rejected goods also require attention because physical arrival alone may not establish an accepted receipt. Investigate exceptions with the receiving and purchasing records, then correct the underlying transaction or obtain a valid credit or revised invoice. Automation is useful for clean matches and consistent exception routing, but it can multiply errors if inaccurate receipts or master data pass through unchecked.

Worked example: A purchase order authorizes 100 units at $20 each, or $2,000. Receiving confirms only 95 acceptable units, supporting 95 × $20 = $1,900 in goods value. The invoice claims $2,000, a five-unit or $100 difference. Check whether five units were missed, rejected or never shipped; resolve that fact before approving the full invoice.

Watch the distinction: Matching the invoice to the purchase order alone overlooks what arrived. An authorized quantity is a purchasing limit, not evidence that every unit was delivered and accepted.

Reference: [1]

Manage Internal Operations and Inventory

23. Build a feasible master schedule

The master schedule converts an aggregate production plan into specific finished items and completion periods. For each period, project available inventory as beginning inventory plus scheduled receipts minus demand. Carry the ending balance forward, and examine a negative balance as a warning that the dated plan cannot satisfy expected demand. Then test whether proposed receipts can actually be produced: materials, resource capacity and lead times may force a different quantity or date. A positive inventory projection alone does not prove the schedule is feasible. Keep forecasts and firm customer orders visible so the same demand is not counted twice. Near-term changes may also disrupt released work, so assess their operational cost before revising dates. The schedule becomes the dated input for material planning and order promising.

Worked example: Period one begins with 40 finished units. A scheduled receipt of 100 and demand of 120 leave 40 + 100 − 120 = 20 units. Period two needs 70 units and has no existing receipt, so its projected balance would be 20 − 70 = −50. Schedule at least 50 units for period two, then verify material and capacity availability.

Watch the distinction: An aggregate plan of 100 units does not specify which item or when it is available. Customer promises depend on the dated master schedule and its feasibility.

Reference: [1]

24. Explode material requirements

Material requirements planning translates dated finished-goods needs into dated component needs. Start with the master schedule, multiply each finished item by the quantity of each component in its bill of material, and account for components needed at lower assembly levels. Subtract usable on-hand stock and scheduled receipts to find net requirements; observe lot-size rules where they apply. Then offset each planned receipt by the component's procurement or production lead time to determine its release date. This sequence matters because a correct total quantity can still arrive too late. Check unit conversions, scrap assumptions and inventory accuracy before trusting the result. Shared components must be summed across all parent items requiring them, while a component already allocated to another order is not freely available.

Worked example: A finished-goods schedule calls for 200 units in week 10. Each needs two brackets, so gross demand is 200 × 2 = 400 brackets. With 50 usable brackets on hand and no other receipts or demand, net need is 350. Assuming an exact-size order and a two-week lead time, release 350 brackets in week 8 for receipt in week 10.

Watch the distinction: Multiplying the bill of material gives gross need, not automatically the order quantity. Usable stock, scheduled receipts, other allocations and lead time affect the planned order.

Reference: [1]

25. Check the capacity constraint

Capacity checks ask whether the dated work plan fits the resources available when the work must occur. Multiply each planned quantity by its required processing time, add setup and other relevant loads, and compare the total with effective available hours. Effective capacity may be lower than calendar hours because of breaks, maintenance, yield loss or staffing. Locate the overloaded period and resource before committing a delivery date. Possible responses include approved overtime, an alternate qualified route, subcontracting, changing lot timing or negotiating a later date. Each has a cost and may move the constraint elsewhere. A machine-hours calculation alone is insufficient if tooling, materials or skilled labor are also limiting. Recheck the schedule after the chosen response so orders and promised dates remain aligned.

Worked example: A work center offers 400 effective hours this week. An order for 250 units needs two hours per unit, or 500 hours, before any additional setup time. The overload is 500 − 400 = 100 hours. If approved overtime supplies 100 effective hours and materials are ready, the order fits; otherwise shift 50 units at two hours each or use qualified alternate capacity.

Watch the distinction: A material-ready order is not necessarily producible on time. Ignoring work-center load turns a feasible component plan into an infeasible customer promise.

Reference: [1]

26. Place inventory at the right stage

Inventory protects a particular interval in the flow, so its location matters as much as its quantity. Raw materials help prevent production from waiting for inputs. Work in process can protect a downstream operation from upstream variation, although too much may conceal a bottleneck and lengthen flow time. Finished goods can cover customer demand during production and replenishment time. To choose a buffer, compare the customer's required response time with the time needed to complete each remaining stage. Place stock after the stages that cannot be completed within the promise, while considering forecast error, customization and carrying cost. A make-to-order product may need component stock rather than finished goods if customer choices determine the final form. Measure service improvement and inventory investment together.

Worked example: Customers require shipment within one day, while production takes five days even when raw material is ready. Adding raw stock leaves the five-day conversion delay intact. Finished goods of the right item can support the one-day promise, assuming picking and transport are available. If the item is customized, holding a suitable intermediate assembly may work only if final configuration takes no more than one day.

Watch the distinction: More inventory anywhere does not produce the same service result. A buffer placed before a five-day production stage cannot satisfy a one-day shipment promise by itself.

Reference: [1]

27. Set a reorder point

A reorder point triggers replenishment when the inventory position reaches expected demand during lead time plus a chosen safety buffer. Inventory position is usable on-hand stock plus firm incoming stock minus backorders or committed quantities, so it differs from what a shelf count shows. Use matching time units: daily demand multiplied by lead time in days. Safety stock protects against variation in demand or replenishment, but its size needs a service and cost decision rather than an arbitrary percentage. Recalculate the point when the demand rate, lead time, reliability or intended service changes. A reorder point tells you when to order; an order-quantity rule determines how much. For intermittent or strongly seasonal demand, a single steady-rate trigger may misrepresent near-term needs.

Worked example: Average demand is 20 units per day, replenishment takes five days and the chosen safety stock is 30 units. Expected lead-time demand is 20 × 5 = 100 units, so reorder when inventory position reaches 130 units. If lead time becomes seven days with everything else unchanged, the point becomes 20 × 7 + 30 = 170 units.

Watch the distinction: Using on-hand stock alone can trigger an unnecessary order when stock is already inbound. Use inventory position, while checking that inbound quantities and dates are dependable.

Reference: [1]

28. Size a safety buffer

Safety stock absorbs uncertainty during replenishment, especially unexpected demand or late supply. Begin with the exposure period and the pattern of variation, then decide what shortage risk the business is willing to carry. More stock generally improves availability but raises carrying, obsolescence and cash costs. An economic comparison needs consistent periods and incremental effects: compare the added cost of the proposed buffer with the shortage cost it is expected to avoid. Avoid treating a service target as a guarantee, because actual demand and lead time can exceed the assumed range. Also ask whether shorter lead time, better forecasts or more reliable suppliers could reduce exposure with less stock. Pooling demand across locations can sometimes lower total buffer needs, provided service and transfer times still work.

Worked example: Suppose 40 extra units cost $2 per unit per year to carry, adding 40 × $2 = $80 annually. Planners estimate that this buffer will avoid two shortages per year, each costing $60, for an expected $120 benefit. Net estimated gain is $40 per year. The recommendation depends on those shortage and demand estimates and excludes any obsolescence cost.

Watch the distinction: A larger safety buffer is not automatically economical. It protects service only against uncertainty within its reach, and its annual benefit must be compared with annual carrying and loss costs.

Reference: [1]

29. Dispose of aging stock

Aging inventory needs a disposition decision based on future use and recoverable value, not its original purchase cost. Check whether demand remains credible, whether the item can be reworked or used elsewhere, and whether it is safe and legal to sell. Compare the net cash from markdown or resale after handling and channel costs with reuse, recycling and disposal alternatives. Include continued storage, damage and obsolescence risk if waiting for a possible sale. Original cost is generally sunk for this forward-looking choice, although accounting treatment may require a separate write-down. A markdown can also displace sales of a current item, so its apparent recovery may overstate its benefit. Record the disposition and adjust planning parameters to avoid buying more of an item already being cleared.

Worked example: One hundred obsolete units can be sold for $4 each, bringing in $400. Handling costs $50, leaving $350 net cash. Scrapping instead costs $40 and yields nothing, a net cash result of −$40. The sale is $390 better on these facts, assuming the units are saleable and the markdown does not reduce profitable sales of a current product.

Watch the distinction: Waiting to recover the original unit cost can deepen a loss when demand has vanished. Compare realistic future proceeds and costs, not the historical price paid.

Reference: [1]

30. Maintain traceability and custody

Traceability connects a product's identity to its origin, processing and destinations. Record lot or serial numbers at receipt, transformation, storage, picking and shipment, including splits and combinations. Chain of custody adds who held or transferred the goods and when, which helps verify handling and locate affected stock. The required detail should match the consequence of an incident: a lot record may suffice for a batch issue, while a unique serial can locate a specific unit. During a suspected defect, first identify the affected identifier, then reconcile the total received or produced with stock on hand, work in process, shipments and returns. Isolate accessible units and use destination records for outreach. Incomplete records widen the population that must be treated as suspect.

Worked example: A suspect lot contains 500 units. Records show 200 still in storage and 300 shipped to two customers, so 200 + 300 = 500 units are accounted for. Hold the 200 accessible units and use lot-linked shipment records to identify the customers and shipped quantities. An item-level balance without lot identity could not separate these units from unaffected production.

Watch the distinction: Knowing that 500 units of an item exist somewhere is not lot traceability. Without movement and destination records, you cannot isolate the suspect batch precisely.

Reference: [1]

31. Use counts to correct inventory records

A cycle count tests whether the recorded quantity agrees with a controlled physical count without requiring a full shutdown. First verify the item, location and unit of measure, then freeze or account for transactions occurring during the count. Calculate the variance and review recent receipts, picks, transfers and adjustments before correcting the record. The direction of the difference can provide a clue: physical stock below the record may indicate an unposted shipment, mispick, shrinkage or wrong location. Correct both the quantity and the transaction cause; otherwise the error will recur. Count frequency can reflect item value, movement and operational risk, while a repeated pattern may justify changing scan points or training. Accuracy matters because planning uses the record, not the stock that happens to be on a shelf.

Worked example: The system shows 240 units, while a controlled count finds 228. The shortage is 240 − 228 = 12 units, or 12 ÷ 240 = 5% of the recorded balance. A recent 12-unit shipment without a posted pick would explain it. Verify that transaction and location, post the missing movement if confirmed, then recount to establish the corrected 228-unit balance.

Watch the distinction: Changing the record to match the count closes the arithmetic difference but may leave the cause intact. Investigate transaction timing and handling before treating the adjustment as a fix.

Reference: [1]

32. Plan maintenance supplies

Maintenance, repair and operating supplies support equipment rather than becoming part of the saleable product. Their inventory decision depends on the consequence of being unavailable, replacement lead time and demand uncertainty. A rarely used part can still deserve a spare if its absence stops a critical line and replenishment takes longer than the tolerated outage. For common, readily available supplies, a simple reorder rule may be sufficient. Check interchangeability, shelf life, storage conditions and whether a repairable part can be returned to service. Coordination with maintenance plans helps distinguish scheduled consumption from random failures. Compare the cost of holding a spare with the likely disruption it prevents, using explicit assumptions about failure and downtime. Do not use annual consumption alone as the measure of importance.

Worked example: A machine needs one seal to restart, and a replacement takes three days to obtain. One spare costs $40; a day of downtime is estimated at $2,000. If the seal fails while no spare is available, the purchase delay could expose several days of production. Holding one spare is readily justified if failure during its usable life is plausible, though the exact expected saving needs a failure estimate.

Watch the distinction: A part used only once in several years is not necessarily low priority. Its criticality depends on the outage caused by its absence and how fast it can be replaced.

Reference: [1]

33. Diagnose performance against plan

Performance against plan needs both an output view and a cost view. Start by comparing actual volume, timing, quality and service with the operational plan. Then compare actual spending with a budget adjusted, when appropriate, for actual activity so higher output is not mistaken for inefficiency. A labor cost variance can arise from higher wage rates, overtime premiums, extra hours per unit or a different product mix. Investigate these causes before recommending cuts: overtime may have protected an important delivery, while rework may show a process problem. Relate the variance to the value of the service achieved and any future effect. The same discipline used in financial planning and analysis helps make a supply-chain variance actionable: quantify it, isolate its driver and choose a response.

Worked example: The plant produces its planned 1,000 units, but actual labor cost is $26,000 against a $22,000 budget for that output. The unfavorable variance is $4,000, or $4,000 ÷ $22,000 ≈ 18.2% of budget. Volume is on target, so inspect hours, rates, overtime and rework before concluding whether the extra spending bought necessary service or reflects inefficiency.

Watch the distinction: Meeting the unit plan does not prove operations performed well. A cost overrun can hide behind target output, while a justified expedite can look unfavorable without service context.

Reference: [1]

34. Choose the improvement method

Choose an improvement method by diagnosing the dominant problem. Lean examines steps that consume time or resources without adding customer value, using flow and waste reduction. Six Sigma studies variation and defects, often through a structured define, measure, analyze, improve and control sequence. Total quality management treats quality as an organization-wide responsibility, including customer focus and sustained process control. Theory of constraints asks which resource limits system throughput, then protects and improves that constraint before expanding other steps. These approaches can work together, but the first intervention should match the evidence. Map the process, measure the loss, identify the governing cause and predict the system effect of a proposed change. A local utilization gain is weak evidence if total output, quality or delivery does not improve.

Worked example: A line's bottleneck can produce 90 units per day, while every other station can handle 120. Increasing one nonbottleneck to 140 leaves line output at 90. If demand is 110 daily, the shortfall is 20 units. Raising verified bottleneck capacity to 110, while keeping other stations at 120, could close that gap if materials and quality remain adequate.

Watch the distinction: Improving the fastest station may make its utilization look better without increasing shipments. Measure the system constraint and its downstream effect before claiming a throughput gain.

Reference: [1]

Manage Supply Chain Logistics

35. Design warehouse roles

A warehouse can hold stock, consolidate or split loads, perform value-added work, or transfer inbound goods quickly to outbound transport. Cross-docking fits predictable flows with matched inbound and outbound timing because it reduces storage and handling time. It still requires accurate advance information, dock capacity and a plan for late arrivals or unmatched quantities. Regional stock can shorten customer response time, but duplicates inventory and adds facility cost. Central storage often pools demand and reduces fixed cost, while increasing some delivery distances. Select the role by comparing required service with total facility, handling, transport and inventory costs. Also check whether the product needs inspection, repacking or temperature control before release. A design that saves storage cost but routinely misses dispatch windows is not a service-feasible design.

Worked example: An inbound load has 100 units, and confirmed outbound orders cover all 100 for trucks arriving at the same time. Cross-docking can route the full load onward without storage if inspection and labels are ready. If the inbound load instead contains 130 units with only 100 assigned, 30 units need a storage, return or later-order plan; the entire load cannot be treated as immediately cross-docked.

Watch the distinction: Cross-docking does not eliminate the need for coordination. Unmatched orders, late trucks or inspection holds can turn an assumed direct transfer into congestion and unplanned storage.

Reference: [1]

36. Select transport by total service cost

Transport decisions combine freight price with transit time, reliability and the cost of inventory tied up in movement. Begin with the required delivery date and destination stock, then eliminate options that cannot meet service without another remedy. Compare feasible modes and carriers on freight, handling, damage risk, expected delay and working capital. A slower mode may be economical for stable replenishment when stock covers the longer transit; an urgent shortage may justify a premium mode for only the exposed quantity. Shipment consolidation can lower cost per unit but may delay an individual order. Check the full door-to-door interval, including pickup and receiving, rather than comparing advertised travel times alone. The best option minimizes total service cost under the customer's actual timing and condition requirements.

Worked example: Air costs $900 and arrives in one day; ground costs $300 and arrives in five. Goods are needed in three days. Without stock or another source, ground is two days late, so air is the feasible choice despite costing $600 more. If destination stock reliably covers those two extra days, ground may meet customer service while saving $600, subject to the stock's carrying cost.

Watch the distinction: Choosing the lowest freight quote can create a costly service failure. Transit time and available coverage determine whether that quote is a feasible alternative at all.

Reference: [1]

37. Control distribution execution

Distribution execution turns inventory records into accurate customer shipments. Receiving verifies identity, quantity and condition; put-away records the storage location; picking retrieves the ordered item and unit of measure; packing checks contents and protection; shipment confirmation records what left and when. A scan at each handoff ties physical movement to the warehouse record and exposes wrong items or quantities before the error propagates. Units of measure deserve special care because cases, inner packs and eaches can share an item code yet represent different quantities. Exceptions should be resolved at the point found, with a corrected transaction and, where needed, a customer update. Timely shipment confirmation also keeps available stock and order status accurate for subsequent planning and promises.

Worked example: An order calls for 12 eaches. A picker takes 12 cases, each holding six units, so the picked quantity is 12 × 6 = 72 eaches, which is 60 too many. Scanning the case unit during picking and comparing 72 eaches with the 12-each order catches the mistake before packing. Correcting the pick restores stock and order accuracy.

Watch the distinction: A matching item code does not guarantee a correct shipment. A case-to-each conversion error can multiply quantity even when the product itself is right.

Reference: [1]

38. Plan cross-border movement

Cross-border planning starts with the product, origin, destination and agreed delivery terms. These facts determine what documents, classifications and controls must be checked and which party arranges and pays for each movement stage. Incoterms allocate specified transport responsibilities, costs and risk of loss between buyer and seller; the sales contract must still address price, payment and other obligations. Build landed cost from purchase price plus buyer-borne freight, insurance, duties, taxes and handling that apply to the stated transaction. Verify the commercial invoice, packing details and transport documents agree on quantity and description, because mismatches can delay clearance. Check classification and admissibility before dispatch. Do not infer a universal import charge from the trade term; actual charges depend on the product and jurisdictions involved.

Worked example: A seller quotes $5,000 for goods. Under the stated agreement, the buyer also pays $700 freight and $300 in import charges. Assuming no other buyer-borne costs, stated landed cost is $5,000 + $700 + $300 = $6,000. Confirm the agreed Incoterm, named place and documents before commitment, since a different allocation could change both responsibility and cost.

Watch the distinction: An Incoterm does not by itself determine every contractual issue or a fixed import charge. Read the named place and agreement, then calculate buyer-borne costs from the actual transaction.

Reference: [1]

39. Route returns to their best use

Reverse logistics collects and sorts goods moving back from customers, then directs each item toward the highest feasible recovery. Verify identity, condition, warranty status and safety before selecting restock, repair, refurbishment, resale, parts recovery, recycling or disposal. Compare net value after testing, labor, transport and channel costs, along with the time required to recover it. A waste hierarchy favors extending useful life when safe and practical, but the highest nominal resale price may not be achievable for every return. Segregate suspect goods so they cannot re-enter saleable stock prematurely. Record the reason for return as well as the disposition: repeated failures may signal a design, packaging or fulfillment problem whose correction is more valuable than processing each return efficiently.

Worked example: A returned unit can be repaired for $30 and sold for $80, leaving $50 before testing, transport and selling costs. Recycling would yield $12. Repair has a $38 larger stated recovery, so choose it if inspection confirms safety and the resale price is realistic. If extra repair-channel costs exceed $38, recycling could become the better financial option.

Watch the distinction: The highest resale price is not automatically the best disposition. Repair, testing, transport and safety requirements can consume the apparent advantage or rule out resale.

Reference: [1]

Manage Customer and Supplier Relationships

40. Translate the voice of the customer

The voice of the customer is useful when translated from broad statements into observable service requirements. Ask what outcome the customer values, how it is measured, when it is measured and what failure costs them. Reliability may mean delivery on the promised date, complete quantities, accurate documents or consistent condition; average transit speed alone does not reveal any of these. Segment customers when their needs differ, and agree on a measure with a clear numerator, denominator and time window. Examine both the average and the spread of outcomes, because a faster average can coexist with frequent late deliveries. Use customer feedback alongside order and shipment evidence to identify the process causing misses. Requirements analysis is useful here because a testable service requirement guides a concrete design decision.

Worked example: A customer says dependable arrival matters more than speed. Of 100 comparable orders, current on-time delivery is 90%, or 90 orders. A proposed process raises it to 98%, meaning eight more orders arrive on time. Another option cuts average transit from three days to two but leaves ten orders late. The first option better meets the stated need, assuming the promise dates are meaningful.

Watch the distinction: A lower average transit time does not prove more reliable service. If late orders persist, the customer's stated requirement remains unmet despite a faster typical shipment.

Reference: [1]

41. Design differentiated service

Service should reflect what a customer needs and what it costs to deliver. Segment customers or orders by requirements such as response time, delivery reliability, order size and handling needs. Then define a service offer and price for each segment. Estimate cost to serve using picking, transport, inventory, returns and support costs, including costs that a faster promise shifts elsewhere in the network. Compare the customer's willingness to pay or the value of retaining the business with that incremental cost. Capacity matters too: promising scarce same-day slots to every order can make the service unreliable for customers who need it most. Review segment results because a valuable customer may have a costly order pattern that calls for a redesigned service, not simply a discount. The CSCP content outline includes customer service offerings within relationship management.

Worked example: A distributor's same-day option costs $25 per order; two-day delivery costs $8. The incremental cost is $25 − $8 = $17 per order. If one customer values same-day receipt at $24, the option creates $7 of value beyond its added cost. If another values it at $10, two-day service is the better offer, assuming both meet their stated needs.

Watch the distinction: Giving every high-revenue customer the fastest service can waste capacity. Check each customer's actual delivery need and the incremental cost of meeting it before setting the promise.

Reference: [1]

42. Define customer-facing measures

A customer-facing measure needs an event, a population and a clock. For on-time, in-full delivery, decide whether the unit is an order, order line or shipment; what quantity counts as full; and which promised date is authoritative. The numerator is the number meeting both conditions, while the denominator is the eligible population in the same period. An order that is complete but late fails, as does an on-time partial order. Agree how cancellations, customer-requested date changes and split deliveries are treated before reporting results. Preserve the original promise as well as approved changes so a late order cannot be made to appear timely by quietly resetting its date. Pair the rate with counts and reasons for failure. Otherwise, a good percentage may conceal a small but costly group of missed orders.

Worked example: Assume 100 eligible orders share the same reporting period. Ninety-two arrive by their promised dates and 90 arrive complete. Records show that 86 meet both tests, so on-time, in-full performance is 86 ÷ 100 = 86%. The average of the separate rates is 91%, but it does not measure orders satisfying both conditions. Six are timely but incomplete, and four are complete but late.

Watch the distinction: Averaging on-time and in-full percentages does not produce the joint rate. Count orders that satisfy both conditions using one agreed denominator and the same promise rules.

Reference: [1]

43. Resolve feedback at its cause

Customer feedback is useful when it can be tied to an order, product, process step and failure mode. First restore service: replace missing goods, correct the invoice or explain the recovery date. Then separate the immediate remedy from corrective action. Examine order entry, inventory records, pick instructions, scans, packing and carrier evidence to locate where the outcome diverged from the requirement. Ask whether the cause is a one-off execution lapse or a repeatable design flaw. A process change should have an owner and a measurable check, such as the rate of missing accessories in subsequent orders. Verify that the change works across relevant sites or shifts. A complaint count alone is a weak measure because customers differ in whether they report a problem; operational checks can reveal defects before another complaint arrives.

Worked example: Ten customers report missing accessories, and ten replacements settle those orders. A packing audit then shows that the accessory is absent from the pick instruction for this product. The team adds the accessory to the instruction and checks the next 50 orders; all 50 contain it. That supports the fix, though later monitoring is needed to see whether the error recurs.

Watch the distinction: Sending replacements closes the immediate customer cases, but it does not remove the cause. Trace the defect through the fulfillment steps and verify a process correction on later orders.

Reference: [1]

44. Share demand with partners

Orders placed by a retailer may reflect batch sizes, promotions, inventory corrections and fear of shortages as well as consumer demand. If each partner forecasts from its immediate customer's orders, fluctuations can grow upstream, a pattern called the bullwhip effect. Sharing point-of-sale sales, inventory on hand, promotion calendars and planned changes helps partners distinguish real demand shifts from ordering noise. Collaborative planning, forecasting and replenishment works only when partners agree which forecast is shared, who authorizes replenishment, how exceptions are resolved and when data are refreshed. The supplier can plan steadier production from consumer consumption, while the retailer retains visibility into stock targets and service promises. Shared data still need checking: late sales feeds, unrecorded returns or stockouts can make apparent sales understate true demand.

Worked example: Retail sales are 100 units in each of four weeks, but store orders alternate between 60 and 140 units. Average orders are still 100 units per week: (60 + 140 + 60 + 140) ÷ 4 = 100. With a stable agreed stock target, the supplier can plan around 100 units of weekly consumption, then add or subtract documented inventory corrections rather than copying each order swing into production.

Watch the distinction: A volatile order stream does not prove volatile consumer demand. Check point-of-sale sales and inventory changes before increasing upstream capacity or safety stock in response to the swings.

Reference: [1]

45. Set up supplier management

Supplier management intensity should follow the supplier's effect on continuity, quality, cost and future capability. Start by identifying whether alternatives exist, how long qualification takes, what capacity is committed and which products depend on the supplier. A transactional supplier may need clear specifications and routine performance reviews. A supplier supporting a critical or unique component needs shared forecasts, capacity reviews, change control and tested continuity arrangements. Set mutual goals: delivery and quality targets for the supplier, timely forecasts and engineering decisions for the buyer. Name escalation contacts and decide what information each side will share. Partnership does not mean surrendering independent verification. Audit evidence and performance trends still matter, especially if apparent good results depend on expediting or unusually high buffers that hide a fragile process.

Worked example: A sole-source component supports products representing 40% of a manufacturer's revenue. The supplier can make 1,000 units weekly, and the buyer forecasts 950. Utilization is 950 ÷ 1,000 = 95%, leaving only 50 units of weekly headroom. A joint capacity and continuity review should test surge options and recovery plans; a quarterly price check would miss this exposure.

Watch the distinction: Calling a supplier strategic because spend is high overlooks dependency. A modestly priced component can warrant close management when qualification is slow or its failure stops a major product line.

Reference: [1]

46. Build a supplier scorecard

A supplier scorecard translates shared expectations into observable results. Define each measure before collecting data: delivery against which requested or confirmed date, quality by defective units or affected lots, and cost by price variance or total delivered cost. Sustainability measures may include verified declarations, audit findings or agreed improvement milestones. State the reporting period, denominator, data source and owner so the supplier can challenge errors and act on valid results. Use trends and exception details alongside an overall score. Weighted averages help compare tradeoffs when measures can compensate for one another, but a mandatory specification or declaration is a gate: good delivery cannot offset its failure. Review leading signals, such as overdue corrective actions, as well as outcomes. The aim is a specific improvement discussion, not a number detached from operating decisions.

Worked example: A supplier delivers 95 of 100 orders on time and 98 of 100 accepted lots meet quality requirements, yielding 95% delivery and 98% quality. Its required material declaration is missing. Even if an illustrative average of the two percentages is 96.5%, the declaration remains an open requirement. The buyer should address it separately and assess affected shipments before relying on the composite score.

Watch the distinction: A strong composite score can conceal a critical failure. Keep mandatory requirements visible as separate gates, then use weighted measures to discuss performance among suppliers that meet those gates.

Reference: [1]

47. Audit supplier capability

An audit tests whether a supplier's system can keep producing acceptable results, not just whether recent deliveries passed inspection. Work backward from important risks: product specifications, process controls, trained operators, calibrated measurement equipment, traceability, change approval and management response to defects. Sample records and observe the process to see whether written procedures match actual practice. Financial and governance evidence may reveal continuity concerns even when product quality is good. Distinguish a finding from its possible consequence; overdue calibration creates uncertainty about measurements, but it does not by itself prove every shipped unit defective. Agree containment or retesting when evidence warrants it, then require a corrective plan with dates and verification. Prioritize findings by the potential effect on product conformity and delivery, rather than counting all audit observations as equal.

Worked example: A supplier shipped 100 units and found no defects during its final check. The gauge used for a critical dimension is overdue for calibration. The 100 accepted units are an output result, but their measurement evidence is less reliable. The buyer asks for calibration and, if appropriate, remeasurement with a verified gauge before treating the clean lot as proof of sustained capability.

Watch the distinction: A defect-free sample is not proof that the underlying process is controlled. Review measurement, change and corrective-action records to judge whether the result can be repeated.

Reference: [1]

48. Manage supplier corrective action

Corrective action follows a sequence because each step answers a different question. Containment asks which material might be affected and keeps it from reaching customers. A remedy replaces, repairs or credits known failures. Root-cause analysis then asks what process condition created the defect and why controls failed to catch it. The supplier changes that condition, documents the change and verifies effectiveness with later production evidence. Verification should match the failure mode: subsequent lots, measurement records or a controlled process trial may be more informative than a promise of retraining. The buyer also checks whether the fix creates a new risk, such as slower capacity or a changed specification. Close the case only when affected goods are accounted for and evidence shows the process is stable; monitor afterward for recurrence.

Worked example: A supplier isolates 50 suspect units and replaces them, satisfying immediate orders. Inspection finds a worn fixture that shifts a critical dimension. The supplier replaces the fixture, checks the first 30 units made with it and examines the next two lots. If results meet the agreed specification and the fixture-maintenance control is updated, there is evidence for closing the corrective action.

Watch the distinction: Replacing defective units is containment and remedy, not proof of prevention. A valid corrective action identifies the process cause and shows through later output that the fix holds.

Reference: [1]

49. Govern shared commitments

Shared plans need rules for commitment because forecasts, requests and confirmed orders carry different obligations. State the forecast horizon, which periods are firm, who can approve changes and how capacity or material costs are handled. When a customer request exceeds confirmed capacity, quantify the gap before promising a date. Then test options such as phased delivery, overtime, alternate sourcing or a revised product mix, including their cost and feasibility. Escalate early enough for the supplier and customer to choose among them. Record the decision in both partners' systems so procurement, production and customer service work from the same quantities. A collaborative forecast improves preparation, but it is not itself proof that labor, materials or transport are available. Good governance turns new information into an explicit revised commitment.

Worked example: A customer increases its monthly request from 800 to 1,100 units. Confirmed supplier capacity is 900, so the request exceeds that capacity by 1,100 − 900 = 200 units. If the supplier can deliver 900 this month and 200 early next month, the customer must accept that phased plan before the buyer confirms the new promise. Any added-capacity option needs a separate feasibility and cost check.

Watch the distinction: Treating the latest customer forecast as a confirmed supply commitment can create an impossible promise. Compare it with available capacity and obtain agreement on the gap and revised dates.

Reference: [1]

Manage Supply Chain Risk

50. Identify risk across the chain

Risk identification starts with the flow that could fail: material, information or funds. Supply risks include unavailable inputs or supplier noncompliance; demand risks include cancellations and unexpected mix changes. Operational risks affect conversion or fulfillment, while financial risks can interrupt a partner's ability to trade. Regulatory changes and external events may cross several stages. Name the initiating event, affected node, time to impact and likely consequence rather than recording only a vague label such as 'delivery risk.' Map dependencies such as a shared port, single data provider or common sub-tier supplier. One event can produce several consequences, and different events can produce the same late shipment. The distinction matters because the owner and response differ: alternate transport helps a port closure, while factory fire requires production recovery or another source.

Worked example: A port closure delays 400 inbound components, a factory fire stops assembly, and a customer cancels 100 finished units. All three can reduce shipments or revenue, but their first responses differ. Logistics checks alternate routes and stock for the port event; operations assesses backup assembly capacity; demand planning revises the finished-goods plan for the cancellation. One generic expediting action cannot resolve all three.

Watch the distinction: Labeling every threat 'supply disruption' hides where it begins. Identify the event, affected flow and first point of impact so the right owner can choose a suitable response.

Reference: [1]

51. Prioritize exposure

Prioritization considers likelihood, loss magnitude and how quickly a disruption would affect service. For a simple monetary comparison, expected annual loss equals annual event probability times loss per event, assuming the estimates and periods are comparable. This screens candidates for analysis, but it does not describe the worst case or a firm's ability to survive it. A rare event with a catastrophic loss may deserve action even when its expected value is modest. Add time to impact, recovery time, concentration and existing controls to a risk register. Separate gross exposure from residual exposure after current safeguards. Test assumptions: probabilities may change with season or location, and correlated events can disable several suppliers at once. Ranking should lead to a decision about investigation or response, not merely a colored heat map.

Worked example: Under hypothetical annual estimates, Risk A has a 20% chance of a $500,000 loss, so expected annual loss is 0.20 × $500,000 = $100,000. Risk B has a 5% chance of a $3 million loss, giving 0.05 × $3,000,000 = $150,000. B ranks higher by expected loss, and its larger single-event impact also merits a separate resilience check.

Watch the distinction: A lower probability does not automatically mean a lower priority. Compare impact and time to effect, and check whether a rare loss could exceed the organization's recovery capacity.

Reference: [1]

52. Choose a risk response

Choose a response by comparing what it changes with what it costs. Avoidance removes an exposure, perhaps by discontinuing a hazardous route or design. Mitigation lowers the chance or consequence through dual sourcing, inventory or process controls. Transfer shifts specified financial consequences through insurance or contract terms, though operational harm may remain. Acceptance keeps the residual risk when further action costs more than its expected benefit or the risk fits agreed tolerance. Quantify gross expected loss, response cost and residual expected loss on the same time basis. Then examine limits that arithmetic misses: correlated failures, delayed backup activation, contract exclusions and customer consequences. Assign an owner, trigger and test for the chosen response. A plan has little value if the alternate source cannot actually produce compliant goods when needed.

Worked example: A disruption has an estimated expected annual loss of $100,000. A backup arrangement costs $30,000 per year and reduces expected annual loss to $40,000. Its expected reduction is $100,000 − $40,000 = $60,000; after the fee, the expected net benefit is $30,000 per year. Choose it only if the backup is qualified and those estimates fit the firm's risk tolerance.

Watch the distinction: Buying insurance or naming a backup does not erase operational exposure. Check what loss is covered, how quickly replacement supply starts and what residual customer impact remains.

Reference: [1]

53. Build continuity around recovery time

Continuity planning compares the likely interruption with the maximum time customers or operations can tolerate. Define the point at which service fails, the time to detect and decide, and the time needed for an alternative to produce usable output. A backup that takes three days to activate cannot cover a one-day tolerance without accessible inventory or another bridge. Convert the uncovered interval into units using a stated demand rate, then check location, shelf life, allocation and transport time. Capacity may ramp gradually, so a simple daily average can understate the gap. Test the plan against scenarios that affect both primary and backup sites, and assign authority to release stock or switch production. Recovery targets should be linked to actual customer promises rather than chosen only for technical convenience.

Worked example: Assume customers can wait two days, the main site will be down four days, and demand is a steady 100 units per day. Days three and four create an uncovered interval of two days, requiring 2 × 100 = 200 accessible units or equivalent backup output. If the backup produces only 50 units daily on those days, stock must cover the remaining 100 units.

Watch the distinction: A four-day outage does not always require four days of buffer. Subtract the delay customers can tolerate, then account for the actual ramp-up of accessible backup capacity.

Reference: [1]

54. Protect information continuity

Information continuity protects the decisions that keep goods moving when systems or data become unavailable or untrustworthy. Identify critical records such as open orders, inventory balances, shipment status, supplier contacts and approved bank details. Limit write access by role, preserve recoverable copies and test restoration so a backup is usable rather than merely stored. A manual fallback needs a cutoff time, unique transaction identifiers and an authority to approve shipments. During recovery, reconcile manual actions with restored records before normal automation resumes; otherwise duplicate shipments, invoices or payments may follow. Data integrity incidents require special care because a system can remain online while displaying altered quantities or destinations. Cross-check high-impact changes against an independent source and communicate which records are trusted. Rehearsals reveal whether staff can act within the service window.

Worked example: An order system fails for one day, but a verified export lists 80 open orders. The warehouse manually ships 25 urgent orders and records their unique order IDs and quantities. After restoration, it compares those 25 IDs with the live system, marks completed shipments and investigates discrepancies before releasing the other 55. This prevents the urgent orders from being fulfilled twice.

Watch the distinction: Restoring the application does not complete recovery. Reconcile transactions performed during the outage, and check the integrity of records before automated picking or payment resumes.

Reference: [1]

55. Monitor leading risk signals

A leading risk signal is useful when it gives enough time to act before a customer failure. Examples include increasing supplier lead time, falling confirmed capacity, overdue quality actions or inventory coverage approaching replenishment time. Define the signal, data source, review frequency and action trigger. The trigger should reflect the response lead time: alerting only at stockout is too late if an alternate source needs weeks to qualify. Separate signal from noise by checking persistence and cause, but do not demand perfect certainty before protecting service. Assign an owner to investigate, decide and document the action. After an event, compare when the signal appeared with when action began and whether the mitigation worked. This feedback helps adjust thresholds and retire indicators that do not predict meaningful exposure.

Worked example: A supplier's confirmed lead time rises from four to six weeks while usable inventory covers five weeks at the forecast demand rate. The new lead time exceeds coverage by 6 − 5 = 1 week. The planner checks open orders and initiates expediting or an approved alternate source now. Waiting until inventory reaches zero would forfeit the remaining five weeks of response time.

Watch the distinction: A dashboard alert is not a mitigation. Link each trigger to an owner, an action and enough lead time for that action to protect the affected orders.

Reference: [1]

Evaluate and Optimize the Supply Chain

56. Reassess supply chain strategy

A supply chain strategy should express the business promise in network and policy choices. Translate goals such as delivery speed, product variety, cost and resilience into measurable constraints. Then examine whether facility locations, inventory placement, sourcing choices and transport modes can meet them together. Compare alternatives on total cost to serve, including fixed facilities, freight, inventory, handling and failure costs. A cheap local activity may raise system cost or violate a service promise. Distinguish structural decisions, such as opening a regional warehouse, from operational changes, such as reserving faster transport for urgent orders. Test demand scenarios and capacity before selecting a design. Use financial measures alongside reliability and asset use, because a faster network can tie up working capital. Revisit the strategy when customer needs or market geography changes.

Worked example: A centralized network delivers in five days, while the business promises two. A proposed cheaper lease saves $40,000 yearly but leaves transit unchanged, so it cannot solve the service gap. A regional stock option costing $65,000 yearly meets the two-day promise for the affected area. Compare its added cost with customer value and demand before choosing the network change.

Watch the distinction: Optimizing warehouse rent before testing the delivery promise can improve a local cost while preserving the strategic failure. Screen network options against service and capacity requirements first.

Reference: [1]

57. Calculate cash-to-cash time

Cash-to-cash time estimates how long cash is tied up between paying for inputs and collecting from customers. Calculate it as inventory days plus receivables days minus payables days, using comparable periods and consistent accounting inputs. Inventory days reflect how long goods are held; receivables days reflect collection time; payables days represent supplier credit. A reduction can come from faster inventory flow or collection, while longer payables terms also reduce the measure. Interpret each lever's consequences. Cutting inventory too far can cause stockouts, and stretching payments can damage supplier capacity or terms. The measure is an average, so it can hide seasonal peaks and differences between product lines. Pair it with service and supplier health, and use the management accounting guide for related working-capital analysis.

Worked example: Inventory days are 45, receivables days are 30 and payables days are 25. Cash-to-cash time is 45 + 30 − 25 = 50 days. If a process change reduces inventory to 40 days while collection and payment terms remain unchanged, the cycle becomes 40 + 30 − 25 = 45 days. The five-day improvement holds only if service and the other terms remain stable.

Watch the distinction: A shorter cash-to-cash cycle is not automatically better. Check whether the change depends on missed sales, excess supplier pressure or a temporary accounting shift.

Reference: [1]

58. Optimize with binding constraints

Optimization compares feasible alternatives, not just their unit prices. Define decision quantities by site or route, then state demand, capacity and service constraints. A solution must meet every binding requirement: a low-cost site cannot supply beyond its capacity, and a cheap route may miss the promised date. Calculate total cost for each feasible mix, including production, freight, inventory and any fixed cost triggered by using a site. Identify the binding constraint because it explains why the apparent cheapest plan fails and where extra capacity would create value. If demand or costs are uncertain, test several scenarios instead of treating a single optimum as permanent. Also check practical details such as minimum order sizes, product qualification and shared bottlenecks. The best mathematical answer is useful only if those assumptions match operations.

Worked example: Demand is 1,000 units. Site A can supply at most 600 units at $8 each; site B can supply at most 500 at $10 each. With equal service and no other costs, use all 600 from A and 400 from B. Total variable cost is 600 × $8 + 400 × $10 = $8,800. An all-A plan exceeds its capacity by 400 units.

Watch the distinction: Choosing the lowest unit-cost site for every unit ignores capacity. Check feasibility first, then allocate remaining demand among qualified options using total delivered cost and service limits.

Reference: [1]

59. Evaluate sustainability tradeoffs

Sustainability decisions should examine economic, environmental and social effects across the product's life, rather than treating one visible metric as the whole result. Define the functional comparison: the same number of protected deliveries or uses, with equivalent safety and quality. Include material production, transport, washing or repair, return logistics and end-of-life handling. Track relevant measures such as waste, emissions, worker conditions and cost, making units and boundaries explicit. Simple payback can screen an investment, but it omits the timing of later savings and nonfinancial effects. A reusable container may reduce disposable packaging yet add return miles and cleaning water. Test loss rates and expected life because the advantage can disappear if containers are damaged or fail to return. Verify claims using supplier evidence and measured operations.

Worked example: Reusable containers require $20,000 initially and are expected to save $6,000 per year in packaging and disposal, assuming a stable shipment volume. Simple payback is $20,000 ÷ $6,000 per year = 3.33 years. If the containers last five years, gross savings total $30,000 and exceed the initial cost by $10,000 before maintenance, return transport and the time value of money.

Watch the distinction: A short payback does not prove an environmental gain. Measure return transport, cleaning, losses and disposal against the same number of protected shipments before claiming a net benefit.

Reference: [1]

60. Test an improvement before scaling

A pilot tests whether an improvement works under real operating conditions and whether it shifts cost or risk elsewhere. Begin with a baseline and a clear outcome measure, including its denominator and observation period. Choose a comparable pilot population, document what changed and record other factors such as demand mix or staffing. Measure intended benefits together with inventory, labor, error rates and customer effects. Compare absolute counts and rates, since a percentage can improve merely because the mix became easier. Set decision criteria before reviewing results, then investigate exceptions and repeatability across sites or seasons. A technology pilot also needs data quality, user adoption and recovery checks; a dashboard that displays earlier warnings has value only if teams act on them. Scale after confirming the benefit can survive normal volume and ownership.

Worked example: Before a visibility pilot, 12 of 100 orders arrive late, a 12% late rate. During the pilot, 5 of 100 are late, a 5% rate, so the observed improvement is 7 percentage points. If average inventory also rises from $200,000 to $250,000, that is a 25% increase. Estimate its carrying cost and check order mix before deciding whether to expand the pilot.

Watch the distinction: A better service percentage alone does not establish that the pilot should scale. Compare like-for-like orders and measure the inventory, labor and system costs used to obtain it.

Reference: [1]

Official references and further reading

ascm-cscp5-current-ecm

All study guides

Forecast and Manage Demand

1. Read demand history carefully

2. Scan demand outside the business

3. Shape demand across the product life cycle

4. Choose a forecasting method

5. Measure error and bias

6. Reconcile demand and supply in S&OP

Manage the Global Supply Chain Network and Information

7. Design for the required service

8. Map product, information and funds

9. Build usable end-to-end visibility

10. Protect connected supply chain systems

11. Govern master data

12. Design a balanced performance view

Source Products and Services

13. Make or buy by total cost and capability

14. Specify sourcing needs and timing

15. Segment purchasing categories

16. Analyze the supply base

17. Influence design before sourcing

18. Qualify a supplier

19. Compare delivered sourcing cost

20. Negotiate workable contract terms

21. Choose a purchase-order approach

22. Reconcile the purchase-to-pay record

Manage Internal Operations and Inventory

23. Build a feasible master schedule

24. Explode material requirements

25. Check the capacity constraint

26. Place inventory at the right stage

27. Set a reorder point

28. Size a safety buffer

29. Dispose of aging stock

30. Maintain traceability and custody

31. Use counts to correct inventory records

32. Plan maintenance supplies

33. Diagnose performance against plan

34. Choose the improvement method

Manage Supply Chain Logistics

35. Design warehouse roles

36. Select transport by total service cost

37. Control distribution execution

38. Plan cross-border movement

39. Route returns to their best use

Manage Customer and Supplier Relationships

40. Translate the voice of the customer

41. Design differentiated service

42. Define customer-facing measures

43. Resolve feedback at its cause

44. Share demand with partners

45. Set up supplier management

46. Build a supplier scorecard

47. Audit supplier capability

48. Manage supplier corrective action

49. Govern shared commitments

Manage Supply Chain Risk

50. Identify risk across the chain

51. Prioritize exposure

52. Choose a risk response

53. Build continuity around recovery time

54. Protect information continuity

55. Monitor leading risk signals

Evaluate and Optimize the Supply Chain

56. Reassess supply chain strategy

57. Calculate cash-to-cash time

58. Optimize with binding constraints

59. Evaluate sustainability tradeoffs

60. Test an improvement before scaling

Official references and further reading

Frequently Asked Questions

Related Study Guides

Association of Chartered Certified Accountants Qualification (ACCA)

Certified Anti-Money Laundering Specialist (CAMS)

Certified Bank Auditor (CBA)