ROI Analysis

ROI (return on investment) analysis calculates and compares the return a decision or initiative produces relative to what it costs. Its job is to give a decision-maker a comparable number across competing options — which project, which vendor, which initiative is worth funding — not to produce a single impressive-looking figure in isolation.

The basic formula

ROI = (net benefit − cost) ÷ cost, usually expressed as a percentage. Net benefit is the value the investment produces; cost is everything spent to produce it. The formula is simple; almost all of the real work in ROI analysis is in defining "benefit" and "cost" honestly and consistently enough that the resulting number actually means something.

Time horizon and discount rate

An ROI figure is only comparable to another ROI figure calculated over the same time horizon. A three-year ROI and a one-year ROI aren't directly comparable even if the percentages look similar, because a longer horizon has more time to accumulate benefit. When comparing options with different horizons, or when future benefit is uncertain or delayed, apply a discount rate so future value is expressed in present terms — a dollar of benefit five years out is worth less than a dollar today, and ignoring that overstates the appeal of longer-horizon options.

What to include as cost

Cost is easy to undercount by stopping at the upfront, visible price tag. A complete cost figure includes ongoing and maintenance costs that continue after launch — support, licensing renewals, staff time, infrastructure, the vendor relationship's Service Level Agreement obligations — not just the initial purchase or build cost. An initiative that looks like a strong ROI on upfront cost alone can look much weaker, or negative, once its full ongoing cost is counted.

Common pitfalls

• Ignoring costs that continue after launch — counting only the upfront investment and omitting maintenance, support, or renewal costs systematically overstates ROI.
• Comparing ROI figures calculated over different time horizons as if they were equivalent — a higher percentage over a longer horizon isn't necessarily the better option once compared on the same basis.
• Treating a rough estimate as more precise than the underlying assumptions justify — presenting a single decimal-precision ROI number when the input assumptions are themselves rough estimates creates false confidence; a range is more honest than a false point estimate.
• No discount rate applied when comparing options with different timing — makes a slower option with distant benefits look more attractive than it actually is in present-value terms.
• Benefit defined inconsistently across options being compared — if one option's "benefit" includes soft, unquantified value and another's doesn't, the comparison isn't actually apples to apples.

Learn more

• Budget Forecasting for the ongoing, org-wide projection of income and expense that a single ROI analysis feeds into rather than replaces.
• Break-Even Analysis for a related calculation focused on the point where cumulative return covers cumulative cost, rather than a ratio.
• Unit Economics for the per-unit cost and revenue figures that often underlie an ROI calculation.
• Feasibility Analysis for the broader evaluation of whether an initiative is viable at all, of which ROI is typically one input.