Project ROI.

Reading the result

Total return is the dollar gain on the money you committed. ROI % is that gain divided by the amount committed. Annualised yield restates it as a per-year rate, which is the only version you can fairly compare against another project, against your borrowing cost, or against simply leaving the money alone.

The rule of thumb for a business decision: a project has to clear your cost of capital, not zero. If a line of credit costs you 9%, a project returning an annualised 6% loses money even though its ROI is positive.

Getting the cost side right

Most ROI calculations are wrong on the numerator, not the denominator. The spend you entered is almost never the full cost of the thing, and the difference is usually large enough to flip the decision.

• Fully loaded labour. If your own team delivered it, their time is a real cost. In Canada an employee costs meaningfully more than salary once CPP, EI, workers’ compensation, employer health tax and vacation pay are counted: see the employee cost calculator for the actual multiple.
• Setup and switching costs. Onboarding, migration, training and the productivity dip while people learn the new thing. These are commonly excluded because they are awkward to attribute, which is exactly why they get spent unaccounted for.
• The tax treatment. An operating expense reduces this year’s taxable income in full. A capital purchase is deducted over years through capital cost allowance instead, so its after-tax cost profile is completely different. Two spends with the same sticker price can have different real costs.
• Opportunity cost. The team that built this did not build something else. ROI cannot see that, and it is frequently the largest number in the decision.

The return side: incrementality is the whole game

For a marketing spend in particular, the honest question is not “how much revenue is attributed to this campaign” but “how much revenue happened because of it that would not have happened anyway.” Those are different numbers and the second is almost always smaller.

Branded search is the classic case: someone who already intended to buy searches your name, clicks the ad, and converts. The platform reports a conversion and a spectacular ROI. The incremental revenue is close to zero, because that customer was arriving regardless. The only reliable way to separate the two is a holdout; deliberately withhold the spend from a comparable group or region and measure the difference in outcomes rather than the difference in attributed clicks.

Also be clear about which margin you are using. Revenue divided by spend is ROAS, an advertising efficiency ratio. ROI uses profit, so it needs revenue less cost of goods and less the spend itself. A 3× ROAS on a product with a 25% gross margin is losing money, and the two metrics being confused for one another is how that goes unnoticed.

When ROI is the wrong tool

ROI has no concept of time value or risk. Over a horizon longer than a year or two, or where the cash arrives unevenly, use a discounted method (net present value or an internal rate of return) because a dollar in year five is not a dollar today. For anything with a real chance of returning nothing, the expected value across outcomes matters more than the ROI of the good case.

The complementary number worth computing alongside ROI is payback period: how many months until the spend is recovered. ROI tells you whether the decision was good; payback tells you whether your cash flow can survive it being good slowly. A project with a strong ROI and a nineteen-month payback can still put a small business under.

Comparing investments rather than business projects? Use investment ROI for total and annualised return on a holding, or the CAGR calculator to compare holdings owned for different lengths of time.