Cost-Benefit Analysis: Steps, Formula, and Examples

By Brian Tr · Updated Oct 10, 2026

Cost-benefit analysis (CBA) is a decision method that puts every cost and benefit of an option in money terms, adjusts future amounts to today's value, and compares the totals. If discounted benefits exceed discounted costs, the option creates value. It's most useful for ranking options that compete for the same budget.

This guide covers the formulas, a seven-step method, a worked example with made-up numbers, a template you can copy, and the places where the method misleads.

Key Facts

• Core idea: Net present value (NPV) equals the present value of benefits minus the present value of costs. Pick the option with the largest positive NPV.
• Two outputs: NPV tells you how much value an option creates. The benefit-cost ratio (BCR) tells you how many dollars of benefit you get per dollar of cost.
• Discounting is required: A dollar next year is worth less than a dollar today, so every future amount gets converted to present value before you compare.
• Government standard: US federal guidance asks analysts to show net benefits at both 3 percent and 7 percent real discount rates, while the UK Green Book sets its social discount rate at 3.5 percent for the first 30 years.
• Ratios can mislead: Both the US and UK guidance warn against picking an option on the highest ratio alone.
• Best for: Investments, build-or-buy calls, system purchases, expansions, and policy choices where benefits can reasonably be counted in money.

What is cost-benefit analysis?

Cost-benefit analysis is a structured way to answer: "Is this worth doing, and is it worth more than the alternatives?" You list every cost and every benefit over a set time horizon, convert them to money, discount them to today's value, and compare.

The US Office of Management and Budget calls benefit-cost analysis a primary tool for regulatory analysis. Its Circular A-4 says that when all benefits and costs can be expressed in money, the method gives decision makers a clear indication of the most efficient alternative, the one with the largest net benefits. The UK's Green Book describes social cost-benefit analysis as the recommended approach for detailed comparison of shortlisted options.

Both documents were written for public spending. The logic transfers directly to a company: replace "society" with "the business" and "social cost" with "total cost of ownership."

The formulas

Net present value (NPV):

NPV = (PV of all benefits) - (PV of all costs)

Present value of a single future amount:

PV = Amount in year t / (1 + r)^t

where r is the discount rate and t is the year.

Benefit-cost ratio (BCR):

BCR = (PV of benefits) / (PV of costs)

A BCR above 1.0 means discounted benefits exceed discounted costs. A BCR of 1.5 means every dollar spent returns $1.50 in value. The BCR and NPV always agree on go or no-go for a single option, but they can disagree when you rank several options of different sizes. That's why most guidance leans on NPV for ranking.

Cost-benefit analysis vs related tools

CBA overlaps with several tools you probably already use. They answer different questions.

A practical way to combine them: use a decision matrix to narrow a long list to two or three options, run cost-benefit analysis on the shortlist, and use scenario planning to stress the winner. The Green Book follows a similar pattern, using CBA for the detailed comparison of a shortlist.

How to do a cost-benefit analysis in 7 steps

Step 1: Define the decision and the alternatives

State the decision in one sentence and list the real options, including "do nothing." The do-nothing case is your baseline. Without it, you end up crediting the project with benefits that would have happened anyway.

Step 2: Set the time horizon

Pick a period that matches the useful life of the investment. Software might get five years, equipment ten. Be consistent across options, or the longer horizon will look better just because it counts more years.

Step 3: List all costs and benefits

Costs are broader than the invoice. Include purchase price, implementation, training, maintenance, internal staff time, and the cost of disruption. Benefits include added revenue, saved labor hours, avoided errors, lower churn, and avoided penalties.

Count only incremental effects, meaning the difference versus the baseline. And watch for double counting: higher revenue and higher profit from the same sale are one benefit, not two.

Step 4: Put a dollar value on each item

Convert hours to dollars using loaded hourly cost, not salary alone. Convert avoided errors to the average cost of an error. When you can't price something, don't force a number. Measure it in its natural units or describe it, and carry it to Step 7.

Step 5: Choose a discount rate and discount

The discount rate reflects the time value of money plus risk. Companies commonly use their cost of capital or a hurdle rate set by finance. Public guidance gives anchors: Circular A-4 asks for estimates at real rates of 3 percent and 7 percent, and the Green Book sets 3.5 percent real for years 1 to 30 of an appraisal. Use real rates with real (inflation-adjusted) cash flows, or nominal rates with nominal cash flows. Don't mix the two.

Then compute present value for each year's costs and benefits.

Step 6: Calculate NPV and BCR

Sum discounted benefits, sum discounted costs, subtract for NPV, divide for BCR. Rank the options by NPV, and use the BCR as a secondary check on capital efficiency when budget is limited.

Step 7: Test the result and add the intangibles

Run a sensitivity analysis: change the key assumptions and see whether the answer flips. Then list the benefits and costs you couldn't price and ask whether they could change the ranking. The next two sections cover both.

A worked example (hypothetical)

Suppose a company with a 40-person support team is deciding whether to buy a ticketing and automation system. These numbers are invented for illustration, not benchmarks.

Assumptions:

• Upfront cost in year 0: $200,000 (license, setup, training).
• Annual net benefit in years 1 to 4: $90,000 (saved agent hours plus fewer escalations, after subtracting ongoing fees).
• Discount rate: 7 percent. Time horizon: 4 years.

PV of benefits is $304,849, so BCR = 304,849 / 200,000 = 1.52. The simple payback period is about 2.2 years ($200,000 / $90,000).

The decision is a "yes" on these assumptions. But a yes on one set of assumptions isn't a decision yet. It's a hypothesis to test.

Sensitivity analysis on the same example

Sensitivity analysis changes one input at a time and watches the result. The Green Book asks analysts to calculate switching values, meaning the value a key assumption would need to reach before an option stops being worth it. Circular A-4 asks for a sensitivity analysis that shows whether results hold under plausible changes to the main assumptions.

Two readings follow. First, the switching value for annual benefit at 7 percent is about $59,046, roughly 34 percent below the base case. Below that, the project loses value. Second, the combined bad case flips the sign, so the decision depends on benefits arriving close to plan. That tells you where to spend management attention: on proving the benefit estimate before signing.

Handling intangible benefits and costs

Not everything fits a spreadsheet. Brand reputation, employee morale, strategic flexibility, and customer trust are real, and they're hard to price.

Both reference documents give the same advice. The Green Book says to monetise where possible, then quantify what you can't monetise, then describe the rest qualitatively. Circular A-4 asks analysts to present a summary table listing unquantified benefits and costs, ranked by importance, and to run a threshold or break-even analysis when they could matter. The threshold question is: "How small could the value of the non-quantified benefits be before the project would yield zero net benefits?"

You can use that in a business setting. If the NPV is -$20,000 and the project also improves customer trust, ask whether trust is plausibly worth more than $20,000 over the horizon. If the answer is clearly yes or clearly no, the decision is easy. If it's murky, that's where your judgment, and maybe a gap analysis of where you stand today, belongs.

Template: cost-benefit worksheet

Copy this into a spreadsheet and fill one sheet per option.

Below the table, add four cells: Total PV of benefits, Total PV of costs, NPV, BCR. Then a short list titled "Unpriced items" with each intangible, its likely direction, and its rough importance.

Cost-benefit analysis by business function

If a decision is really about owning versus renting a capability, pair this method with the build, borrow, or buy framework, which frames the options and leaves CBA to price them.

Do's and don'ts

Do

• Include a do-nothing baseline. Benefits only count if they exceed what would happen anyway.
• Count opportunity cost. Staff time and cash committed here can't be used elsewhere.
• Use the same horizon and discount rate across options. Otherwise the comparison isn't fair.
• Show your assumptions. Circular A-4 says a good analysis is transparent enough that a qualified third party could see how you reached each estimate.
• Rank by NPV, then sanity-check with BCR. NPV shows scale. BCR shows efficiency per dollar.
• Haircut the benefits. The Green Book describes optimism bias as the proven tendency for appraisals to be over-optimistic and advises raising cost and duration estimates while lowering benefit estimates.

Don't

• Don't pick on the highest ratio alone. Circular A-4 states that the ratio of benefits to costs is not a meaningful indicator of net benefits, and the Green Book says UK government doesn't choose projects solely on the highest BCR. A small project with a 4.0 ratio can create less total value than a large one with a 1.4 ratio.
• Don't include sunk costs. Money already spent doesn't change which option is best from here.
• Don't count transfers as benefits. A cost moved from one team's budget to another's creates no new value. The Green Book notes that including transfers can push the BCR artificially toward one.
• Don't double count. Revenue and the margin on that revenue are one benefit.
• Don't hide the intangibles. An unpriced factor that you leave out of the document doesn't stop being a factor.
• Don't treat the result as a verdict. A CBA is a structured argument. If the answer is close, say so.

Where cost-benefit analysis falls short

CBA works best when benefits and costs are measurable and the future is reasonably predictable. It struggles when most of the value is strategic or hard to price. Circular A-4 itself says BCA is less useful, and can even be misleading, when important benefits and costs can't be expressed in money.

It also hides distribution. A project with positive NPV may still shift costs onto one team or customer segment while another captures the gain. And the answer is only as good as the inputs: small changes to the discount rate or the benefit estimate can reverse the result, which is why the sensitivity step isn't optional.

When uncertainty is the main issue, build two or three futures first with scenario planning, then run the CBA under each. For choices with several competing non-financial criteria, a decision matrix fits better. For choices that sit at the top of the organization, decision-making for CEOs covers the wider habits that surround the numbers.