Cost-Benefit Analysis: A Step-by-Step Method

Cost-benefit analysis is a method for deciding whether an action is worth taking by comparing everything it costs with everything it delivers, expressed in money and adjusted for the timing of each cash flow. It is the logic behind most business cases, capital approvals and policy decisions. Done well, it forces you to list costs and benefits explicitly, to put values on them, to discount the future and to test how sensitive the answer is to what you assumed. Done badly, it is a justification exercise in which the benefits are generous and the costs are partial. This guide sets out the steps, works through a five-year example with checked arithmetic including net present value and the benefit-cost ratio, shows a sensitivity test, covers intangible items and lists common errors. It is part of business analytics for finance and operations, follows how to write a business case and the ROI calculator template, and leads to the ROI formula and how to calculate ROI for a business project. It supports the decisions made in budgeting and forecasting.

The steps

1. Define the decision and the options. What is being decided, and what are the alternatives, including doing nothing?
2. Set the scope and horizon. Who is affected, and over what period? Use the useful life of the investment.
3. List the costs. Direct and indirect, one-time and recurring, including internal time and opportunity costs.
4. List the benefits. Cash savings, additional margin, avoided losses and, separately, intangible benefits.
5. Put a monetary value on each item and state the source and the confidence.
6. Discount to present value at a rate that reflects the cost of capital and the risk.
7. Compare: NPV, benefit-cost ratio and payback.
8. Test sensitivity, and identify the assumptions that decide the answer.
9. Decide and document.

A worked example

A distributor considers a customer ordering portal (hypothetical). The portal costs $120,000 to build and launch, and $20,000 a year to run. It is expected to bring benefits of $65,000 a year from fewer phone orders, fewer order errors and faster payment, for five years. The discount rate is 8 percent.

The annuity factor for five years at 8 percent is (1 - 1.08^-5) / 0.08 = (1 - 0.68058) / 0.08 = 3.9927.

• Present value of running costs: 20,000 × 3.9927 = $79,854
• Total present value of costs: 120,000 + 79,854 = $199,854
• Present value of benefits: 65,000 × 3.9927 = $259,526
• Net present value: 259,526 - 199,854 = $59,672
• Benefit-cost ratio: 259,526 / 199,854 = 1.30

Check: 65,000 × 3.9927 = 259,525.5, and 20,000 × 3.9927 = 79,854; 259,526 / 199,854 = 1.2986, which rounds to 1.30. A ratio above 1.0 means the present value of benefits exceeds the present value of costs, so the portal creates value above the 8 percent required return. Each dollar of cost buys $1.30 of benefit in present value terms.

For comparison, the undiscounted figures are total costs of 120,000 + 5 × 20,000 = $220,000, total benefits of 5 × 65,000 = $325,000, a net gain of $105,000 and an ROI of 47.7 percent. Discounting reduces the net gain to $59,672 and shows that the project is less attractive than the headline suggests, though still positive.

Sensitivity: how wrong can the benefits be?

The benefits are the weakest part of the case. If they turn out to be 20 percent lower, at $52,000 a year, the present value is 52,000 × 3.9927 = $207,620, the NPV is 207,620 - 199,854 = $7,766 and the benefit-cost ratio is 1.04. The project barely survives. If benefits are 25 percent lower, at $48,750 a year, the present value is $194,644 and the NPV is -$5,210, a loss. The break-even level of annual benefit is 199,854 / 3.9927 = $50,055, which is about 23 percent below the base estimate (1 - 50,055 / 65,000 = 0.23). The honest message to the decision maker is that the portal is worth building if the benefits come in within about 23 percent of the plan, and the team should focus on the evidence for the biggest benefit before committing.

Sensitivity should cover costs as well. If the build cost overruns by 25 percent, to $150,000, total present value of costs rises to $229,854, and the NPV falls to 259,526 - 229,854 = $29,672, still positive. The result is more sensitive to benefits than to the build cost, which tells you where to put your effort.

Handling intangible benefits

Some benefits are real but hard to value: better customer experience, brand effects, employee morale, reduced risk of a rare event. Three approaches are common.

1. Leave them out of the numbers and list them as upside. This is the safest approach and the best default.
2. Estimate a value using a proxy, such as the cost of customer churn avoided, and flag it as an estimate with a low confidence.
3. Use a break-even test. Ask how much of the intangible benefit would be needed for the project to break even, and judge whether it is plausible. In the example, if the NPV were -$20,000, the intangible benefits would need to be worth about $5,000 a year in present value terms (20,000 / 3.9927).

Do not assign a number to every soft benefit to make the total look bigger. Reviewers become skeptical when the intangibles dominate the case.

Whose costs and benefits count?

Define the perspective. A business case normally takes the company's perspective, but some decisions affect customers, suppliers or employees in ways that matter, such as a policy that shifts cost onto suppliers who then raise prices. Be explicit about who is included. Be careful of transfers: a saving to one department that is a cost to another is a transfer, not a benefit to the company.

Choosing the discount rate

Use the company's cost of capital as a base, and add a premium for riskier projects. Use the same rate for all projects you compare. Test the result at a rate a few points higher and lower: if a project is attractive only at a low rate, it is fragile. For short projects of a year or less, discounting has a small effect and can be skipped.

Common mistakes

1. Omitting costs, especially internal time, maintenance and the cost of change.
2. Double counting benefits, for instance counting both time saved and the cost of the people whose time was saved.
3. Counting transfers as benefits.
4. Ignoring timing, by adding up undiscounted flows over long periods.
5. Selective scope: including all the benefits but only part of the costs.
6. Sunk costs counted as costs of the new decision. Money already spent should be ignored.
7. No sensitivity analysis.
8. Using the analysis to justify a decision already made.

Comparing two options

Cost-benefit analysis is most useful when it compares real alternatives. Suppose the distributor's other option is to keep taking orders by phone and email but add a part-time order coordinator at $28,000 a year, which brings benefits of $34,000 a year from fewer errors and quicker processing (hypothetical). With the same five-year, 8 percent assumptions:

• Present value of costs: 28,000 × 3.9927 = $111,796
• Present value of benefits: 34,000 × 3.9927 = $135,752
• NPV: 135,752 - 111,796 = $23,956
• Benefit-cost ratio: 135,752 / 111,796 = 1.21

The portal has the larger NPV ($59,672 against $23,956) and the higher ratio (1.30 against 1.21). To see what the extra investment buys, look at the difference between them. The portal costs $88,058 more in present value (199,854 - 111,796) and brings $123,774 more in present value of benefits (259,526 - 135,752). The incremental NPV is 123,774 - 88,058 = $35,716, and the incremental benefit-cost ratio is 123,774 / 88,058 = 1.41. The extra spending on the portal is justified on its own merits, not merely carried along by the cheaper option. That is the incremental logic that every comparison of options should use. See incremental ROI.

Presenting the results

Show the decision maker a one-page summary: the options, the present value of costs and benefits for each, NPV, benefit-cost ratio, payback, and the sensitivity result in a sentence such as "the portal remains attractive unless benefits fall more than 23 percent below plan." Put the list of costs and benefits, with sources and confidence levels, in an appendix. Separate what is measured from what is estimated, and what is cash from what is a judgment. A reader who can see how the answer was built is more likely to trust it, and more likely to spot an error that you can fix before the decision.

When a simpler method will do

Not every decision needs a five-year discounted model. For a small purchase with a payback under a year, a simple comparison of annual cost and annual benefit is enough. For decisions that are required by law or safety, the question is the cheapest effective way to comply. Use the full method when the amounts are large, the benefits arrive over several years, the options differ in timing or the decision will be reviewed by people who were not part of it. Match the effort to the stakes.

FAQs

What is cost-benefit analysis?

A method for deciding whether an action is worthwhile by comparing the monetary value of its costs and benefits, adjusted for timing.

What is a benefit-cost ratio?

The present value of benefits divided by the present value of costs. A ratio above 1.0 means benefits exceed costs.

How do I value intangible benefits?

Keep them outside the numbers as upside, use a proxy with clear caveats or use a break-even test to judge how large they would need to be.

What discount rate should I use?

Your cost of capital, adjusted for risk, and applied consistently across the projects being compared.

What is the difference between cost-benefit analysis and ROI?

ROI expresses net gain as a percentage of cost, usually without discounting. Cost-benefit analysis discounts cash flows and compares total benefits with total costs, giving NPV and a benefit-cost ratio.