Financial Appraisal and Design Decision-Making

Introduction

Warehouse design is often presented as a physical planning problem: determine storage capacity, select material-handling equipment, position receiving and dispatch areas, and define flows. In practice, however, every design is also an investment decision. A warehouse layout commits capital, affects operating costs for many years, constrains future expansion, and influences service performance, safety, labour requirements, energy use, and technological capability.

Financial appraisal therefore provides a disciplined way to compare competing designs. It does not replace operational judgement. Rather, it makes assumptions visible and allows decision-makers to evaluate whether a technically attractive design creates sufficient economic value under plausible future conditions.

A sound appraisal addresses five questions:

1. What resources must be committed initially and over the design life?
2. What operating benefits and costs will each alternative generate?
3. How does the timing and risk of cash flows affect value?
4. How robust is the preferred alternative when assumptions change?
5. What implementation path preserves flexibility as demand and technology evolve?

The analysis is most useful when finance, operations, engineering, information technology, procurement, and occupational safety specialists develop it jointly.

Learning objectives

By the end of this chapter, the reader should be able to:

• distinguish capital expenditure, operating expenditure, cash flow, and accounting cost;
• construct a life-cycle total cost of ownership model for warehouse alternatives;
• calculate and interpret ROI, payback period, NPV, and related measures;
• compare automation, conventional, and hybrid layouts under uncertainty;
• conduct sensitivity, scenario, and risk analysis;
• explain real options in phased warehouse investment;
• prepare an evidence-based business case and technology roadmap;
• identify governance, ethical, and implementation risks in financially driven design decisions.

1. The economic architecture of a warehouse design decision

A warehouse alternative should be defined as a complete operating system, not merely a drawing. The appraisal unit may include building modifications, storage media, conveyors, picking equipment, warehouse management software, control systems, scanners, robotics, utilities, training, commissioning, maintenance, and changes to staffing. Comparing a highly automated design with a conventional design while omitting software, integration, or maintenance produces a biased result.

1.1 Capital expenditure and operating expenditure

Capital expenditure (CAPEX) is expenditure that creates or improves an asset expected to provide benefits over multiple accounting periods. Typical warehouse CAPEX includes:

• building construction or alteration;
• racking, mezzanines, docks, and safety barriers;
• conveyors, sortation, automated storage and retrieval systems, and vehicles;
• control software, servers, sensors, and systems integration;
• design, installation, testing, and commissioning;
• initial training and start-up inventory, where relevant.

Operating expenditure (OPEX) is incurred in running the system. Examples include labour, energy, consumables, equipment maintenance, software subscriptions, insurance, space rent, damage, and periodic compliance inspections.

The classification is not always universal. Accounting standards, tax law, and organisational policy may treat software, leases, development costs, or implementation services differently. Financial appraisal should therefore distinguish the economic substance of a cost from its accounting classification.

1.2 Relevant and incremental cash flows

Only differences caused by the decision should normally enter an alternatives analysis. If an existing building will be used under every option, its unavoidable historical cost is a sunk cost and should not affect the choice. If one alternative requires a larger building, the incremental rent or construction cost is relevant.

Relevant cash-flow categories include:

• initial investment at time 0;
• incremental annual operating costs and savings;
• additional working-capital requirements;
• taxes and tax depreciation effects, when the appraisal is after tax;
• residual or salvage value;
• decommissioning, removal, and restoration costs;
• avoided losses, such as reduced damage or fewer service failures.

A benefit should be monetised only when there is a credible causal relationship. For example, increased throughput has financial value only if the organisation can sell more, avoid outsourcing, or reduce a binding constraint. A theoretical capacity increase is not automatically a cash benefit.

1.3 Cost behaviour and operational drivers

Warehouse costs are often driven by activities rather than by floor area alone. A useful cost model separates:

• fixed costs, such as equipment depreciation, rent, and system licences;
• variable costs, such as picking labour, packaging, and energy per movement;
• step costs, such as adding a supervisor or shift when volume crosses a threshold;
• failure-related costs, such as downtime, rework, and expedited transport.

If annual volume is Q_t units, cost per unit is c_t, and annual fixed cost is F_t, a simplified annual operating-cost model is:

Here O_t is total operating cost in year t, measured in currency per year. The model assumes a linear variable-cost relationship over the relevant range. In real warehouses, labour productivity, congestion, overtime, and step costs may make the relationship nonlinear.

2. Total cost of ownership

Total cost of ownership (TCO) estimates the complete economic burden of owning and operating an alternative over a defined period. It is especially valuable when an option has a high purchase price but lower labour, energy, or error costs.

A basic undiscounted TCO is:

where:

• I₀ = initial investment;
• O_t = operating cost in year t;
• M_t = maintenance cost;
• U_t = other incremental costs, such as training or insurance;
• R_t = major replacement or refurbishment cost;
• S_n = residual value at the end of year n;
• n = analysis horizon.

All terms must be expressed in the same currency and price basis. TCO is not a discounted valuation: a cost occurring in year 8 is treated as equal to a cost occurring today. This makes TCO easy to communicate but incomplete for investment decisions.

A discounted TCO uses present values:

where r is the annual discount rate. The discount rate must be consistent with inflation and tax assumptions. Nominal cash flows, which include expected inflation, require a nominal discount rate; real cash flows require a real discount rate.

Worked example: comparing two layouts

An organisation is comparing Layout A, a conventional selective-rack system, with Layout B, a goods-to-person automated system. The analysis is illustrative and uses constant annual cash flows before tax.

Layout A has annual cost of $1,140,000; Layout B has annual cost of $860,000. Ignoring discounting, the undiscounted TCOs are:

TCO_A = 1,200,000 + 10(1,140,000) − 100,000 = $12,500,000

TCO_B = 2,400,000 + 10(860,000) − 300,000 = $10,700,000

Thus, Layout B has an undiscounted ten-year advantage of $1,800,000. This conclusion depends on the assumptions: in particular, the automated system must achieve the forecast labour saving, remain available, and not require an unplanned major upgrade.

3. Investment appraisal methods

No single metric is sufficient. A robust business case presents at least NPV, payback, and operational measures such as capacity, service level, safety, and resilience.

3.1 Return on investment

A simple project ROI can be expressed as:

The numerator and denominator must use a clearly defined period and cost basis. Some organisations use average annual net benefit divided by initial investment, while others use accounting profit divided by invested capital. These definitions are not interchangeable. ROI is useful for communication but can ignore timing, risk, and benefits after the selected period.

3.2 Payback period

The payback period is the time required for cumulative net cash flows to recover the initial investment. If annual net cash flow is constant, initial investment is I₀, and annual net cash flow is C, then:

This shortcut applies only when cash flows are constant and there is no need to account for discounting. With unequal cash flows, cumulative cash flow should be tracked year by year. If recovery occurs part-way through a year:

Payback is intuitive and useful for liquidity risk, but it ignores cash flows after payback and generally ignores the time value of money.

3.3 Net present value

NPV measures the value created in today’s currency:

where CF_t is net cash flow in year t, S_n is residual value, and r is the discount rate. If taxes, working capital, and disposal costs are relevant, they must be included in CF_t or the terminal amount.

A positive NPV indicates that the project is expected to earn more than the required return represented by r, subject to the quality of forecasts and risk adjustment. When mutually exclusive alternatives use the same service requirement, the alternative with the higher NPV is normally preferred, provided non-financial constraints are satisfied.

Worked NPV example

Suppose an automation project requires $2,400,000 initially and generates net cash savings of $280,000 in each of years 1–3, $360,000 in each of years 4–7, and $420,000 in each of years 8–10. It has a residual value of $300,000 in year 10. At a 10% discount rate:

• present value of years 1–3 savings ≈ $696,700;
• present value of years 4–7 savings ≈ $924,800;
• present value of years 8–10 savings ≈ $323,600;
• present value of residual value ≈ $115,700.

Total present value is approximately $2,060,800. Therefore:

NPV = 2,060,800 − 2,400,000 = −$339,200

On these assumptions, the project does not meet a 10% required return. The result is not evidence that automation is inherently uneconomic; it indicates that the specific investment, timing, savings, and risk assumptions require revision or that an alternative should be selected.

3.4 Internal rate of return and equivalent annual cost

The internal rate of return (IRR) is the discount rate that makes NPV equal to zero. It can be useful, but it may mislead when cash flows change sign more than once, because multiple IRRs can exist. NPV should remain the primary value criterion.

When alternatives have unequal lives, equivalent annual cost (EAC) can support comparison. For a present value of cost PVC, discount rate r, and life n:

This capital-recovery factor assumes regular annual periods and a constant discount rate. It is appropriate when alternatives provide equivalent service and can be repeated or replaced under comparable conditions. If lives cannot reasonably be repeated, a common-horizon or explicit replacement analysis is preferable.

4. Building an evidence-based business case

A business case should show how operational evidence becomes financial value. A recommended structure is:

1. Decision statement: what approval is requested and by when?
2. Baseline: current volume, labour hours, service performance, error rate, space use, and constraints.
3. Alternatives: include a credible do-minimum option, not only the preferred technology.
4. Requirements: throughput, storage, peak demand, service levels, safety, resilience, and compliance.
5. Assumptions register: source, owner, date, confidence, and sensitivity of each assumption.
6. Financial model: CAPEX, OPEX, cash flow, tax, working capital, residual value, and discount rate.
7. Non-financial evaluation: safety, scalability, maintainability, workforce effects, sustainability, and strategic fit.
8. Risk and sensitivity analysis: identify failure conditions and mitigations.
9. Implementation plan: stage gates, procurement, commissioning, training, and benefits realisation.
10. Recommendation: explain why the preferred option dominates or best balances value and risk.

Multi-criteria decision analysis

Financial value is necessary but not always sufficient. A weighted scoring model can incorporate safety, service, flexibility, strategic alignment, and environmental performance. If alternative j receives score s_ij on criterion i, and criterion weight is w_i, the weighted score is:

Scores are meaningful only when criteria are defined, scales are consistent, and weights are agreed transparently. A weighted score should not be used to conceal a regulatory failure or unsafe design. Some criteria should be treated as mandatory gates rather than tradeable points.

5. Sensitivity, scenario, and risk analysis

Forecasts in warehouse design are uncertain. Demand, wage rates, energy prices, equipment availability, system performance, and implementation duration may all differ from the base case.

5.1 One-way sensitivity analysis

One-way sensitivity changes one assumption while holding others constant. Common variables include volume growth, labour saving, CAPEX, uptime, maintenance cost, discount rate, and residual value. A tornado chart can rank variables by their effect on NPV.

For a variable x, a local sensitivity measure can be represented as:

This is not a probability measure; it indicates exposure to a change in the selected variable. Nonlinear systems require several test points rather than one slope.

5.2 Scenario analysis

Scenario analysis changes several connected assumptions coherently. For example:

• downside: lower demand, delayed commissioning, higher wage inflation, and lower automation availability;
• base: most probable volume and performance;
• upside: stronger demand and successful productivity improvement.

A scenario is not a forecast probability. If probabilities p_s are credibly assigned to scenarios, expected NPV may be calculated as:

Expected value should not be interpreted as a guaranteed outcome. Decision-makers should also examine downside magnitude and the organisation’s ability to absorb it.

5.3 Probability distributions and Monte Carlo analysis

When uncertain variables can be represented by probability distributions, simulation generates a distribution of NPV rather than a single point estimate. Outputs may include the probability that NPV is positive, the 5th-percentile NPV, and the probability of exceeding a payback threshold. The validity of simulation depends on the quality of distributions, correlations, and causal assumptions. False precision is a serious danger.

5.4 Risk-adjusted decision-making

Risk can be treated through a discount rate, scenario probabilities, explicit risk costs, or a combination. Increasing the discount rate is a blunt instrument and may double-count risk already reflected in conservative cash flows. A transparent model identifies specific risks, their likelihood, financial consequence, owner, and mitigation.

A risk register should distinguish:

• risk of demand being lower than forecast;
• technical integration and cybersecurity risk;
• supplier viability and spare-parts risk;
• operational disruption during installation;
• labour availability and industrial-relations risk;
• safety and regulatory risk;
• obsolescence and data-governance risk.

6. Real-options analysis and phased investment

Traditional NPV often assumes that management must follow a fixed plan. In reality, a warehouse project may create options: defer investment, expand capacity, abandon a failing technology, switch suppliers, or stage automation after demand is demonstrated.

A real option has value when:

1. uncertainty is material;
2. management can respond after uncertainty is reduced;
3. the response is not costless; and
4. the initial design preserves that flexibility.

For example, a modular conveyor interface, standard data architecture, and reserved expansion zone may cost more initially but reduce the cost of future expansion. A low-NPV pilot may be justified if it provides reliable information before a much larger commitment. The option value should not be added casually as an arbitrary percentage. It should be estimated through decision trees, scenario valuation, or comparison of staged and immediate strategies.

A decision tree may contain:

• an initial pilot investment;
• uncertain demand or technical performance;
• a later choice to expand, modify, or stop;
• cash flows conditional on each branch.

The value of flexibility is the difference between the value of the flexible strategy and the value of a rigid strategy, using consistent assumptions. Real-options reasoning is particularly relevant to robotics, artificial intelligence, autonomous mobile robots, and rapidly evolving software platforms.

7. Technology roadmaps for warehouse design

A technology roadmap connects current capability with future operational requirements. It should not be a catalogue of fashionable technologies. A robust roadmap identifies the problem, required data, process maturity, architecture, investment gate, and measurable benefit.

A practical sequence is:

Stage 1: Stabilise. Standardise processes, location master data, barcode discipline, inventory accuracy, safety controls, and basic performance measurement.

Stage 2: Digitise. Implement or improve WMS functionality, mobile scanning, dashboards, slotting analytics, and integration with enterprise systems.

Stage 3: Assist. Introduce pick-to-light, voice systems, automated guided or autonomous mobile vehicles, and decision-support tools where the business case is proven.

Stage 4: Automate selectively. Deploy goods-to-person, sortation, robotic picking, or automated storage where volume, SKU profile, labour economics, and service requirements justify it.

Stage 5: Optimise and learn. Use simulation, digital twins, predictive maintenance, and controlled AI applications while maintaining human oversight and cybersecurity.

Each stage should have entry criteria, a budget, measurable benefits, data ownership, and a stop or redesign decision. Technology should be modular where possible, with open interfaces, documented data structures, and exit provisions in supplier contracts.

8. Illustrative case: Northstar Distribution Centre

The following case is fictional and illustrative.

Background and problem

Northstar Distribution serves retailers from a 12,000 m² facility. Annual volume is 3.2 million order lines, with pronounced seasonal peaks. The current layout uses manual pallet storage and cart-based picking. Congestion, overtime, and picking errors are increasing. Management is considering a conventional expansion, a highly automated system, or a phased hybrid design.

Alternatives

• Alternative A: add floor space and selective racking; retain mainly manual picking.
• Alternative B: install a fully automated goods-to-person system immediately.
• Alternative C: implement a hybrid design: improve slotting and WMS first, add modular goods-to-person equipment for fast-moving items, and retain flexible manual capacity for uncertain demand.

The project team establishes common service requirements: 99.5% order accuracy, peak throughput of 1,800 order lines per hour, two-shift operation, and compliance with safety regulations.

Analysis

The team collects twelve months of transaction data, observes travel time, validates SKU velocity classes, and models peak demand rather than using average volume. It identifies that only 35% of SKUs generate 78% of order lines. This supports selective automation rather than automating every product.

The initial financial model shows that Alternative B has the highest CAPEX and a negative base-case NPV because its capacity is underused outside peak periods. Alternative A has the lowest investment but requires more labour and has limited scalability. Alternative C has a moderate initial investment and a second-stage expansion option.

Sensitivity analysis identifies three dominant variables: realised labour productivity, peak volume growth, and automation uptime. Under the downside scenario, Alternative C can defer the second-stage investment and continue using manual capacity. Alternative B cannot easily reduce its fixed cost, while Alternative A becomes constrained by labour availability.

The recommendation is Alternative C, subject to a six-month pilot and stage-gate approval. The pilot must demonstrate inventory accuracy, software integration, achieved pick rates, safe interaction between people and equipment, and a credible maintenance plan. The business case treats the second-stage expansion as conditional rather than assuming it will occur.

Outcome and learning points

The case illustrates five principles:

1. A high nominal capacity does not necessarily create economic value.
2. Peak demand and product characteristics should shape automation scope.
3. Flexibility can have financial value even when it is not visible in a static NPV.
4. Data quality and process stability are prerequisites for technology benefits.
5. A recommendation should specify evidence required before the next investment gate.

9. Governance, ethics, and implementation

Financial appraisal can create harmful incentives if managers reward only short-term cost reduction. Automation may reduce repetitive work but alter employment, skill requirements, surveillance practices, and work intensity. Ethical appraisal should consider consultation, fair transition, reskilling, accessibility, and the distribution of benefits and burdens.

Safety must be a design constraint. A positive NPV cannot justify inadequate guarding, poor pedestrian segregation, unsafe manual handling, or insufficient emergency procedures. Regulatory requirements, building codes, electrical standards, machinery safety rules, employment law, and data-protection obligations vary by jurisdiction and must be verified with qualified specialists.

Technology projects also require cybersecurity and data governance. Warehouse control systems, sensors, cloud platforms, and supplier portals may create attack surfaces. The business case should include secure architecture, access control, patching, backups, incident response, vendor assurance, and continuity arrangements.

Implementation risk is frequently underestimated. A financially attractive design may fail because of poor commissioning, inaccurate master data, inadequate training, or a transition plan that disrupts customer service. Benefits should therefore be assigned to named owners and measured through a benefits-realisation schedule. Typical measures include labour hours per order line, throughput at peak, order accuracy, equipment availability, energy per order, safety incidents, and actual versus forecast NPV drivers.

10. Limitations and best practices

Financial models are representations, not facts. Common limitations include:

• forecasts based on short or unrepresentative demand histories;
• omission of congestion and nonlinear capacity effects;
• double-counting labour savings and revenue benefits;
• inconsistent treatment of inflation, tax, and discount rates;
• ignoring implementation disruption and learning curves;
• overstating residual value for specialised equipment;
• treating uncertain benefits as certain cash flows;
• relying on a single vendor’s performance claims;
• using a weighted score to compensate for an unsafe or noncompliant option.

Best practice is to maintain an auditable model with version control, documented assumptions, independent review, and clearly separated input, calculation, and output sections. Use ranges rather than false point precision. Validate operational assumptions through time studies, simulation, supplier acceptance tests, and pilot data. Present both the preferred option and the conditions under which it would cease to be preferred.

Chapter Summary

Financial appraisal converts warehouse design alternatives into comparable economic and operational propositions. Capital expenditure, operating expenditure, relevant incremental cash flows, residual value, and working capital must be defined consistently. Total cost of ownership reveals life-cycle burden, while NPV incorporates the time value of money and is generally the strongest primary financial criterion. ROI and payback are useful communication and liquidity measures but have important limitations.

A credible business case combines financial analysis with service, capacity, safety, flexibility, sustainability, and strategic criteria. Sensitivity and scenario analysis expose dependence on uncertain assumptions; simulation can estimate distributions of outcomes when probability information is defensible. Real-options analysis recognises that staged investment, modularity, and the ability to defer or expand may have economic value.

Technology roadmaps should progress from process stability and data quality toward digitisation, assisted operations, selective automation, and optimisation. Governance must address safety, employment impacts, cybersecurity, regulatory compliance, supplier dependence, and benefits realisation. The best design decision is not necessarily the most automated or the cheapest initially; it is the option that creates durable value while remaining robust, governable, and adaptable.

Key Terms

Capital expenditure (CAPEX): Investment in assets or capabilities expected to provide benefits over multiple periods.

Operating expenditure (OPEX): Recurring expenditure required to operate a warehouse system.

Relevant cash flow: Incremental cash flow caused by accepting one alternative rather than another.

Sunk cost: A past cost that cannot be changed by the current decision.

Total cost of ownership: The complete life-cycle cost of acquiring, operating, maintaining, and disposing of an asset.

Discount rate: Rate used to convert future cash flows into present value.

Net present value: Present value of benefits and costs, including initial investment and residual value.

Payback period: Time required for cumulative cash flows to recover initial investment.

Return on investment: Ratio of net benefit to a defined investment or cost base.

Internal rate of return: Discount rate that makes project NPV equal to zero.

Residual value: Net value recovered at the end of the analysis horizon.

Equivalent annual cost: Uniform annual cost equivalent to a present value of costs.

Sensitivity analysis: Examination of how an outcome changes when an assumption changes.

Scenario analysis: Evaluation of coherent combinations of assumptions representing possible futures.

Real option: Value created by the ability to defer, expand, contract, abandon, or switch an investment.

Stage gate: A formal decision point at which evidence is reviewed before proceeding.

Benefits realisation: Process of measuring and securing intended project benefits.

Multi-criteria decision analysis: Structured comparison using multiple weighted or mandatory criteria.

Technology roadmap: Time-phased plan linking capabilities, investments, dependencies, and decision gates.

Review Questions

Short Answer Questions

1. Distinguish CAPEX from OPEX using two warehouse examples for each.
2. Why should sunk costs normally be excluded from an alternatives analysis?
3. State two limitations of the payback period.
4. What is the purpose of a residual value in an NPV calculation?
5. Define a real option in the context of warehouse design.

Long Answer Questions

1. Explain how a total cost of ownership model should be constructed for an automated warehouse. Discuss assumptions, cost categories, and limitations.
2. Compare NPV, ROI, payback period, and IRR as tools for warehouse investment decisions.
3. Describe how sensitivity analysis, scenario analysis, and Monte Carlo simulation differ and explain when each is appropriate.
4. Discuss the role of non-financial criteria in selecting among warehouse layout alternatives.
5. Develop a technology roadmap for an organisation moving from manual warehousing toward selective automation.

Analytical / Application Questions

1. A manual layout requires $1.5 million initially and costs $1.25 million annually. An automated layout requires $3 million initially and costs $900,000 annually. Both have ten-year lives; residual values are $100,000 and $350,000 respectively. Calculate and compare undiscounted TCO, then explain why a discounted analysis is also required.
2. A proposed automation project has a base-case positive NPV but becomes negative when labour savings are 20% lower. Identify the operational evidence you would collect before approval and recommend appropriate risk mitigations.
3. Design a stage-gate decision process for a phased warehouse automation project. Specify the evidence required at the pilot, expansion, and full-scale implementation gates.

Multiple Choice Questions with Answers

Question 1

Which measure explicitly incorporates the time value of money in a warehouse investment appraisal?

A. Simple ROI

B. Payback period without discounting

C. Net present value

D. Undiscounted TCO

Correct Answer: C. Net present value

Question 2

Which item is normally a sunk cost in a current warehouse alternatives analysis?

A. Incremental rent for a larger building

B. Historical expenditure on an existing rack system that cannot be recovered

C. Future maintenance cost

D. Additional commissioning cost

Correct Answer: B. Historical expenditure on an existing rack system that cannot be recovered

Question 3

A project requires $1,000,000 and produces constant annual net cash flow of $250,000. Ignoring discounting, its simple payback period is:

A. 2 years

B. 3 years

C. 4 years

D. 5 years

Correct Answer: C. 4 years

Question 4

In the formula NPV = −I₀ + ∑ CF_t/(1+r)^t + S_n/(1+r)ⁿ, what does r represent?

A. Annual throughput

B. Residual value

C. Discount rate

D. Replacement cost

Correct Answer: C. Discount rate

Question 5

Which statement about payback period is correct?

A. It always accounts for cash flows after recovery

B. It necessarily gives the same ranking as NPV

C. It can be useful for liquidity assessment but may ignore later benefits

D. It is identical to IRR

Correct Answer: C. It can be useful for liquidity assessment but may ignore later benefits

Question 6

Which situation most clearly creates a real option?

A. A sunk cost has already been paid

B. A modular design allows management to expand after demand is observed

C. A fixed annual rent is incurred

D. A historical invoice is recorded

Correct Answer: B. A modular design allows management to expand after demand is observed

Question 7

If nominal cash flows include inflation, the appraisal should generally use:

A. A nominal discount rate

B. A real discount rate of zero

C. No discount rate

D. Only the inflation rate as the discount rate

Correct Answer: A. A nominal discount rate

Question 8

What is the main purpose of one-way sensitivity analysis?

A. To prove that a forecast is certain

B. To change every assumption simultaneously

C. To identify how an outcome responds to changes in one assumption

D. To eliminate operational risk

Correct Answer: C. To identify how an outcome responds to changes in one assumption

Question 9

Which criterion should normally be treated as a mandatory gate rather than traded against financial points?

A. Minor aesthetic preference

B. Regulatory safety compliance

C. Preference for a familiar colour scheme

D. Optional meeting-room size

Correct Answer: B. Regulatory safety compliance

Question 10

In a multi-criteria scoring model, weights should normally satisfy:

A. Each weight must exceed 1

B. The weights must sum to 0

C. The weights must sum to 1

D. All weights must be identical

Correct Answer: C. The weights must sum to 1