Return on Investment (ROI)

Finance · 3 min read

ROI is the simplest formula in business and the most often misused. It asks one question: for every dollar I put in, how many came back? The math is easy. Picking what to count is where most people get it wrong.

One dollar in. How many dollars out?

A marketing manager runs a campaign that costs $50,000 and generates $200,000 in sales. She walks into the next meeting and announces a 300% ROI. Everyone nods approvingly. She gets a bigger budget next quarter.

She may be right. She may also be wildly wrong. The number is only as good as what she chose to count — and almost everyone who calculates an ROI quietly chooses to count it in their favor.

Return on Investment is the ratio of the gain from an investment to the cost of the investment, expressed as a percentage. The formula is straightforward: subtract the cost from the gain, divide by the cost, multiply by 100. An investment of $10,000 that returns $12,000 has produced a 20% ROI. An investment of $10,000 that returns $8,000 has produced a negative 20% ROI. The math is high-school arithmetic.

The reason ROI is so popular is that it converts any investment into a single, comparable number. A 20% ROI on a marketing campaign can be compared to a 15% ROI on new equipment, which can be compared to a 25% ROI on hiring a new salesperson. The investments are completely different in nature. The ROIs are not. That portability is what makes ROI the most-used decision tool in business.

It is also what makes ROI so easy to manipulate.

The marketing manager who reported a 300% ROI on her $50,000 campaign was counting the gross sales the campaign generated — $200,000. But the campaign did not produce $200,000 of profit. It produced $200,000 of revenue. If the products sold had a 40% gross margin, the actual gross profit from the campaign was $80,000. Subtract the $50,000 campaign cost and the real gain was $30,000. That is a 60% ROI, not a 300% ROI. Still a great campaign. Just not the one she described.

The four most common mistakes in calculating ROI are all variations of the same problem: counting the wrong things.

Counting revenue instead of profit. Revenue is bigger than profit, so it makes the ROI look bigger. But the dollars that have to be subtracted to find profit do not just disappear — they are real costs that the investment did not actually eliminate.

Ignoring the cost of time. A 20% ROI achieved in one year is dramatically better than a 20% ROI achieved over five years. The simple ROI formula has no concept of time, which is why financial professionals usually pair it with measures like internal rate of return or payback period. A 20% return next month is excellent. A 20% return over a decade is dismal.

Missing the opportunity cost. Every dollar invested in one thing is a dollar not invested in something else. A campaign that produced a 60% ROI sounds great until you realize that the money could have produced an 80% ROI in another channel. ROI in isolation tells you whether an investment was good. ROI in comparison tells you whether it was the best available. (See Opportunity Cost for the longer version.)

Ignoring downstream costs. A new customer acquired by a marketing campaign is not pure gain. They will require support, account management, ongoing service, and they may eventually churn. A more honest ROI calculation incorporates the full customer lifetime value and the full cost of serving that customer over time — not just the cost of acquiring them.

Used carefully, ROI is one of the most powerful comparison tools in business. Used carelessly, it is one of the easiest ways to justify a decision that has already been made. The discipline of calculating ROI honestly — counting all the costs, all the time, all the alternatives — is what separates the strategic operator from the budget defender.

Before believing any ROI number, including your own, ask three questions. What did you count as the gain? What did you count as the cost? What did you leave out? The honest version of the calculation is almost always smaller than the version that gets reported.