APICS Certified in Planning and Inventory Management (CPIM) — Study Guide

APICS Certified in Planning and Inventory Management (CPIM) — Study Guide

Credential: APICS Certified in Planning and Inventory Management (CPIM) Issuing body: ASCM — Association for Supply Chain Management (APICS brand retained on the certification) Jurisdiction: International (US-based) Level: Professional Prerequisites: None — ASCM lists no eligibility requirements Renewal: 75 certification maintenance points every 5 years (100 for CPIM Fellow)

CPIM is the operations-planning credential: it certifies you can run the planning engine of a manufacturing business — forecasting demand, planning supply, scheduling the floor, and controlling inventory — inside an ERP system. It is the natural companion to the CMfgE for manufacturing engineers who own planning.

1. Pathway — How the Certification Is Earned

1. No eligibility gate. Anyone may register — but the exam assumes working familiarity with manufacturing operations and ERP concepts.
2. Study the Learning System / Exam Content Manual. ASCM's CPIM Learning System is the canonical prep; the Exam Content Manual (ECM) defines the tested domains. Most working professionals prepare 3–6 months.
3. Sit the single computer-based exam at a Pearson VUE center or via OnVUE online testing, scheduled on demand.
4. Maintain with 75 points per 5-year cycle (education, industry work, professional activities).

Exam format (verified 2026):

• One exam — CPIM 9.0 consolidated the old Part 1/Part 2 structure into a single comprehensive exam (treat any two-exam study guide as outdated)
• 150 multiple-choice questions: 130 scored + 20 unscored pretest items (answer everything — no penalty for wrong answers)
• 3.5 hours (210 minutes)
• Scaled score 200–350; 300 to pass (do not convert this to a percentage — ASCM never published one; the scale holds the standard across exam forms)
• Typical fee band ~$1,265–$1,755 depending on membership/bundle (verify current pricing at ascm.org before budgeting)

2. Knowledge Domains

Domain A — Supply chain and operations strategy

The planning hierarchy. Business plan → production plan (aggregate, in product families, months) → master production schedule (MPS — end items, weeks) → material requirements plan (MRP — components, time-phased) → production activity control / shop-floor dispatch. Each level disaggregates the one above and operates on a finer time bucket. The exam tests which decisions live at each level: capacity strategy at the production-plan level, order promising at MPS, component timing at MRP, sequencing at PAC.

Push vs pull. MRP is push (plan-driven, forecast/schedule launches work); Kanban/lean is pull (consumption-driven). Hybrid reality: MRP plans, pull executes — the exam tests knowing which logic governs each decision.

S&OP (Sales and Operations Planning). The monthly executive process reconciling demand and supply plans: demand review → supply review → pre-S&OP → executive S&OP. Output: one agreed operating plan. The exam's strategy questions orbit S&OP because it's where planning meets P&L.

Domain B — Demand management and forecasting

Forecasting methods. Qualitative (sales-force composite, customer surveys, Delphi, executive judgment — for new products/no history) vs quantitative: time-series (moving average, exponential smoothing, trend/seasonal decomposition) vs causal (regression on drivers like price or promotions). Know selection logic: stable demand → moving average/exponential smoothing; trend → add trend component (Holt's); seasonality → seasonal indices; new product → qualitative.

Exponential smoothing mechanics. F_{t+1} = αA_t + (1−α)F_t. α near 1 = responsive (chases noise); α near 0 = stable (lags shifts). The exam tests the trade-off, not just the formula.

Forecast error and bias. MAD (mean absolute deviation), MAPE, tracking signal = RSFE/MAD (bias detector — |TS| > 4ish signals systematic bias; the exam tests bias vs random error because bias breaks planning systems). Safety stock exists because of forecast error and lead-time variability — the exam connects them explicitly.

Demand management. Beyond forecasting: order promising (ATP — available to promise: uncommitted MPS quantity; CTP — capable to promise adds capacity/material checks), customer segmentation, and managing the forecast-to-order boundary (make-to-stock vs assemble-to-order vs make-to-order vs engineer-to-order — the decoupling point determines everything about planning).

Domain C — Supply planning: MPS, MRP, capacity

MPS. The master schedule states what end items to build, how many, when (weekly buckets). MPS logic: project available balance, ATP calculation, and the planning time fence (inside the demand time fence, changes need authorization — the exam tests fence discipline because nervous replanning destroys schedules).

MRP mechanics — the exam's computational heart. Inputs: MPS (demand), BOM (structure), inventory records (on-hand, scheduled receipts, lead times, lot sizes, safety stock). Logic per part-period: gross requirements → net requirements (gross − projected available − scheduled receipts) → planned order receipts → planned order releases (offset by lead time). Low-level coding (explode the BOM level by level so each part is planned once at its lowest occurrence). Pegging (trace a component requirement up to its parent — for rescheduling decisions). Lot-sizing rules: L4L (lot-for-lot — minimal inventory, many setups), EOQ/POQ (economic trade), FOQ (fixed quantity). Nervousness: MRP replans everything when inputs wiggle — mitigations: time fences, firm planned orders, safety stock/time.

Capacity planning. CRP (capacity requirements planning — from MRP's planned releases: load vs capacity per work center), rough-cut capacity planning (at MPS level — the quick check), input/output control (monitor queue and WIP). Load > capacity → the plan is fiction: actions are overtime, subcontract, alternate routings, or replanning the MPS (the exam's correct answer is usually to fix the plan, not to wish).

BOM structures. Single-level vs indented; engineering BOM vs manufacturing BOM; phantom BOMs (transient subassemblies — blow through, don't stock); planning BOMs/super-bills (for ATO variety). Scrap/yield factors inflate gross requirements (plan ÷ yield).

Domain D — Inventory management

Inventory functions and costs. Cycle stock (lot-size), safety stock (variability buffer), anticipation/in-transit/hedge stocks. Carrying cost (capital, storage, risk/obsolescence — typically 20–30%/yr of unit value in textbook problems), ordering/setup cost, stockout cost. The exam's cost questions are EOQ-family trade-offs.

ABC classification. A items (~20% of SKUs, ~80% of value — tight control, frequent review), B (moderate), C (simple rules, larger lots). Cycle counting by class (A counted most often) replaces annual wall-to-wall counts.

Reorder-point systems. ROP = demand during lead time + safety stock; safety stock = z × σ_dDLT (z for the service level — 1.65 ≈ 95%, 2.33 ≈ 99%; σ of demand during lead time = σ_d × √LT for constant LT). Independent-demand items (finished goods, service parts) use ROP/EOQ; dependent-demand items (components) use MRP — the fundamental classification, tested relentlessly.

Lean inventory. Kanban quantities, consignment/VMI (vendor-managed inventory — shifts the planning burden contractually), and the waste lens on inventory (overproduction's hiding place).

Domain E — Production activity control, distribution, quality/continuous improvement, technology

PAC. Order release, dispatching rules (SPT — shortest processing time — minimizes average flow time; EDD — earliest due date — minimizes maximum lateness; CR — critical ratio — dynamic priority), bottleneck management (Theory of Constraints: the bottleneck sets throughput — schedule it first, protect it with buffers; non-bottlenecks' "efficiency" is irrelevant), input/output control.

Distribution. DRP (distribution requirements planning — MRP logic applied to the distribution network), warehouse roles, transportation trade-offs (mode selection vs inventory — faster transport substitutes for safety stock).

Quality and continuous improvement. TQM, Six Sigma DMAIC at literacy level (the exam touches it; the shared Six Sigma guide goes deep), lean wastes (TIMWOODS), 5S, Kaizen — the exam expects you to connect quality capability to planning stability (incapable processes inject the variability that safety stock pays for).

Technology. ERP as the system of record (master data discipline — BOM accuracy, inventory accuracy, routing accuracy: the exam's mantra is that MRP output quality = input data quality), APS/advanced planning, demand-driven MRP (DDMRP — decoupling points + buffers positioned strategically; know it as the modern alternative to forecast-driven MRP), and the basics of supply-chain visibility.

3. Key Equations (with units and assumptions)

• Exponential smoothing: F_{t+1} = αA_t + (1−α)F_t. α ∈ (0,1).
• MAD = Σ|A_t − F_t|/n; Tracking signal = RSFE/MAD (RSFE = running sum of forecast errors).
• EOQ: Q* = √(2DS/H). D annual demand (units/yr), S order cost ($), H holding cost ($/unit-yr).
• ROP = d̄ × LT + SS; SS = z × σ_d × √LT (constant-LT case). d̄ in units/day, LT in days.
• MRP netting: NetReq_t = GrossReqt − ProjectedAvailable{t−1} − ScheduledReceipts_t (+ SS floor). Planned releaset = Planned receipt{t+LT}.
• ATP (discrete): ATP = MPS − committed orders (in the period before the next MPS receipt); cumulative ATP sums forward.
• Number of Kanbans = (demand during LT × (1 + safety factor)) ÷ container size.
• Capacity load: load (hours) = Σ(order qty × run time/unit + setup time) per work center per period; compare to demonstrated capacity.
• Inventory turns = COGS ÷ average inventory value; days of supply = 365 ÷ turns.

4. Worked Examples

Example 1 — MRP explosion. End item A needs 2× B; B needs 3× C. MPS: 100 A in week 5. Lead times: A=1, B=2, C=3 weeks. On-hand: B=50, C=100. L4L throughout, no safety stock. Step-by-step: Gross B in week 4 (release A in week 4 for receipt week 5): 200. Net B = 200 − 50 = 150 → planned receipt week 4, release week 2. Gross C in week 2: 150 × 3 = 450. Net C = 450 − 100 = 350 → receipt week 2, release week −1... negative time — infeasible. Lesson: the exam engineers exactly this: the correct answer is to expedite, reschedule the MPS, or release C late with a documented exception — MRP exposes infeasibility; it doesn't fix it.

Example 2 — Safety stock. Daily demand: mean 40 units, σ = 8. Lead time 5 days constant. 95% service (z = 1.65). Step-by-step: SS = 1.65 × 8 × √5 = 1.65 × 8 × 2.236 = 29.5 → 30 units. ROP = 40 × 5 + 30 = 230 units. Lesson: safety stock grows with √LT — doubling lead time multiplies SS by only 1.41, which is why lead-time variability (not just length) is the real enemy.

Example 3 — EOQ vs L4L. Component: D = 5,000/yr, S = $100, H = $5/unit-yr, unit cost $20. Step-by-step: EOQ = √(2 × 5,000 × 100/5) = √200,000 = 447 units (~11 orders/yr). L4L would order exact net needs — minimal holding, maximal setups. Lesson: EOQ wins when setup cost dominates; L4L wins for expensive or bulky items (and is the lean default). The exam asks you to choose with reasoning, not recite.

Example 4 — ATP. MPS receipts of 200 in weeks 1 and 4. Committed orders: 120 (week 1), 60 (week 2), 90 (week 4), 40 (week 5). Step-by-step: ATP week 1 = 200 − 120 − 60 = 20 (covers until next receipt). ATP week 4 = 200 − 90 − 40 = 70. A new 50-unit order wanted week 2: only 20 ATP available → promise 20 in week 2, 30 in week 4 (or check CTP). Lesson: ATP prevents the sales team from selling the same unit twice — the exam's order-promising questions are ATP arithmetic.

Example 5 — Bottleneck scheduling. Work center WC-10 runs at 100% utilization; others at 60–70%. A rush order arrives. Step-by-step: Per TOC: schedule WC-10 first (it's the constraint — its lost hour is a lost system hour); sequence the rush job on WC-10 by earliest due date among bottleneck jobs; non-bottlenecks subordinate (their idle time is fine — running them faster just builds WIP). Lesson: "balance capacity, not flow" is wrong; balance flow to the bottleneck.

5. Standards/Regulations Cross-Reference

• APICS Dictionary / ASCM body of knowledge — the definitional authority the exam is written against; terminology disputes are settled here, not in textbooks.
• ISO 9001:2015 — QMS context for the quality/continuous-improvement domain.
• ISA-95 / IEC 62264 — the ERP↔MES (L4↔L3) interface standard; the technology domain's structural reference.
• Incoterms 2020 — distribution/transport questions reference delivery terms (EXW, FOB, DDP) for landed-cost and responsibility boundaries.
• UFLPA / forced-labor and tariff regimes — the exam's supply-design questions increasingly assume constrained sourcing; the manufacturing good-to-know guide covers BABA and ESPR as the regulatory backdrop.
• CSCP/CLTD (ASCM) — sibling credentials: CSCP for end-to-end supply chain, CLTD for logistics/distribution; CPIM is the internal operations credential. Sequence them by career direction (see the good-to-know guide).

6. Pitfalls & Trap Patterns

1. Dependent vs independent demand. Components → MRP; finished goods/service parts → ROP/EOQ. Misclassifying is the #1 conceptual miss.
2. The scaled score. 300/350 is not 86%. Don't compute a percentage target; aim for domain-level readiness.
3. Pretest questions. 20 unscored items are mixed in — you can't identify them, so answer everything with full effort.
4. MRP nervousness. Questions describing constant rescheduling test whether you reach for time fences and firm planned orders, not faster computers.
5. Bottleneck logic. Any answer maximizing non-bottleneck utilization is wrong under TOC.
6. Safety stock drivers. SS covers variability (demand and lead-time), not average demand — answers adding SS for the mean double-count.
7. Two-exam study guides. CPIM is one exam (9.0). Materials describing Part 1/Part 2 are outdated — adjust your prep accordingly.

Exam strategy: 150 questions in 210 minutes = 84 seconds each — comfortable but not luxurious. The exam is scenario-heavy: read for which domain the scenario sits in (forecasting? MRP? capacity?), then apply that domain's logic. Calculation questions are straightforward if the formulas are automatic — drill EOQ, ROP/SS, MRP netting, and ATP until they're reflexes. No penalty for wrong answers: never leave one blank.

7. Practice Problems

1. MPS needs 300 units of X in week 6. Each X needs 4 Y (LT 2 weeks, on-hand 200). When must Y orders release, and how many (L4L)? Answer: Gross Y needed week 4: 1,200. Net: 1,000. Release week 2 (LT 2).

2. Daily demand mean 100, σ 15, LT 4 days, 99% service (z = 2.33). Safety stock and ROP? Answer: SS = 2.33 × 15 × 2 = 69.9 → 70. ROP = 400 + 70 = 470.

3. A planner runs MRP daily with no time fences and complains of constant schedule churn. Diagnose and prescribe. Answer: Classic MRP nervousness. Prescribe: planning time fence + demand time fence, firm planned orders inside the fence, and safety stock/time instead of replanning — stabilize before optimizing.

4. EOQ inputs: D = 8,000, S = $75, H = $6. Compute Q and total annual ordering + holding cost at Q. Answer: Q* = √(2 × 8,000 × 75/6) = √200,000 = 447. Orders/yr = 17.9; ordering cost = 17.9 × 75 = $1,342; holding = (447/2) × 6 = $1,341 — equal at optimum, as theory promises.

5. Which dispatching rule minimizes average flow time, and which minimizes maximum lateness? Answer: SPT minimizes mean flow time; EDD minimizes maximum lateness. CR is the dynamic compromise for mixed priorities.

6. A C-item with steady demand is on MRP with weekly replanning. What's wrong? Answer: C-items (low value) deserve simple ROP/min-max rules, not MRP's planning overhead. ABC logic: match control intensity to value.

7. Forecast: α = 0.3, last actual 520, last forecast 500. Next forecast? If actuals then run 560, 575, 590 against forecasts ~505–520, what does the tracking signal tell you? Answer: F = 0.3 × 520 + 0.7 × 500 = 506. Persistent positive errors → tracking signal climbs past ±4 → systematic bias (model missing an uptrend) — switch to a trend-capable method, don't just raise α.

Original ETS reference writing, verified against public sources on 2026-09-30. Thin sourcing is flagged inline. Not legal advice — confirm current editions with the issuing body.