Quick answer: return on investment = (gain minus cost) divided by cost, times 100. Buy for $5,000, get back $6,300 net, and ROI is 26%. The number only becomes meaningful once you annualize it and count every cost, the two steps most ROI claims quietly skip.
ROI in five facts:
- Formula: net gain / total cost x 100
- The missing dimension: ROI ignores time; 24% in one year and in six years are different animals
- Annualized fix: (1 + ROI)^(1/years) – 1
- Costs count fully: fees, taxes, maintenance, and your hours belong in the base
- Universal language: the same formula prices stocks, rentals, ad campaigns, and certifications
How to calculate ROI, start to finish
The worked example, done honestly:
| Line | Amount |
|---|---|
| Purchase | $5,000 |
| Sale proceeds | $6,200 |
| Dividends collected along the way | +$150 |
| Trading fees and costs | -$50 |
| Net gain | $1,300 |
| ROI | 26% |
Two habits shown there: income (dividends, rent) joins the gain, and costs subtract before dividing. The ROI calculator runs the arithmetic and the annualizing in one pass.
Enter what you paid, what you got back, and how long it took to get simple and annualized ROI with the comparison charted.
Annualizing: the step that makes ROIs comparable
Raw ROI has no clock, so it cannot rank opportunities. The conversion:
Annualized ROI = (1 + total ROI)^(1 / years) – 1
| Claim | Timeframe | Annualized |
|---|---|---|
| “Up 24%” | 1 year | 24% |
| “Up 24%” | 2 years | 11.4% |
| “Doubled my money” (100%) | 10 years | 7.2% |
| “Up 50%” | 6 months | 125% |
Row three is the humbling one: doubling in a decade is exactly the market’s ordinary 7.2%, as the Rule of 72 predicts. Every impressive-sounding return deserves the question “over how long?” before any admiration is spent.
Counting the costs people forget
ROI inflation is almost always a denominator crime:
- Rental property: the purchase price is only the beginning; closing costs, repairs, vacancies, insurance, property tax, and management hours all belong in the base
- Stocks and funds: expense ratios, trading costs, and taxes on gains and dividends shave the net
- Business and marketing: a “$5,000 campaign returning $20,000” must charge for the staff time, tools, and fulfillment behind the revenue, and count margin rather than revenue
- Your time: 200 personal hours on a $10,000 flip is a hidden wage; price it, or the ROI is partly unpaid labor wearing a costume
What counts as a good ROI?
Context sets the bar, and the bar is your alternative:
| Arena | Typical annualized range |
|---|---|
| Cash (HYSA, CDs) | 3 to 5%, near risk-free |
| Diversified stocks, long run | 7 to 10% with real volatility |
| Rental real estate | Mid single to low double digits, work included |
| Small business / marketing | Wide open; survivors quote big numbers, the failed are silent |
Any pitch promising steady returns far above these bands is describing either extra risk or fiction. A good ROI is one that beats your honest alternative after costs, at a risk you can hold, which is exactly the comparison NPV formalizes with a discount rate.
ROI’s limits, and the right tool beside it
- No time axis: fixed by annualizing, or by NPV and IRR for multi-year cash flows
- No risk axis: a 12% ROI from a lottery-shaped bet and from a boring index are not equals; volatility is a cost paid in sleep
- Averages hide sequences: +50% then -50% averages to zero and loses 25%, the compounding asymmetry from the compound interest guide
- Yield is not total return: a stock’s payout is one component; price change completes it, per the dividend yield guide
Three ROI audits, start to finish
The formula only bites when the denominator is honest, so run three common cases at full cost. The rental flip: purchase and rehab $180,000, sold at $220,000, and the headline screams 22%; add closing costs both directions, six months of carrying costs, and the agent’s commission, call it $18,000 all-in, and the true gain is $22,000 on $198,000, an 11% ROI earned over eight months of weekends, respectable and half the brag. The certification: a $3,000 course produces a $4,000 raise, “133% ROI” in the testimonial; the raise arrives as salary, so the keep-rate math from a typical bracket trims it to roughly $3,000 of take-home per year, a one-year ROI near 100% that then repeats annually, which is the honest and still-excellent version. The marketing campaign: $5,000 of ads “returned $20,000,” except the $20,000 is revenue on 40% margins, so the gain is $8,000 minus the $5,000 spend, a 60% ROI, real but a fifth of the slide-deck number. Same formula each time; the difference was refusing to let costs hide.
The vocabulary around ROI, decoded
ROI travels with cousins that get swapped carelessly. CAGR is annualized ROI with a formal name, the smooth rate that would have produced your lumpy result. Total return is ROI’s investment-flavored twin: price change plus income, the pairing the yield guide insists on. ROAS, beloved of ad dashboards, is revenue over ad spend, which is not ROI at all until margins and overhead convert revenue into profit; a 4x ROAS on 20% margins is a money-losing campaign wearing a trophy. And “10x return” language from startup land describes multiple-on-invested-capital with no clock attached, which is why the annualizing step exists: 10x over 12 years is 21% a year, spectacular; 10x over 30 years is 8%, an index fund with extra steps and much worse sleep.
Making ROI a habit instead of a headline
The formula’s highest use is prospective, not retrospective. Before any significant outlay, write down the expected ROI and its ingredients: what it costs all-in, what it should return, by when, and what would count as failure. The certification, the ad campaign, the rental, the new equipment, each gets one honest line in a decision journal. Then, on the anniversary, score it against reality. The scoring does two things no spreadsheet can: it calibrates your own optimism (most people discover a personal inflation factor within three entries), and it converts vague regret or vague pride into specific, reusable lessons about where your estimates leak. Investors and businesses that keep this loop quietly compound judgment the way accounts compound dollars, and judgment, unlike returns, carries no market risk.
Ranges beat points: ROI under uncertainty
Every prospective ROI is a guess wearing confidence, so make the guessing explicit. Price three scenarios before committing: the base case with honest numbers, a downside where revenue disappoints and costs bloat by a third, and an upside you would sheepishly admit to hoping for. A campaign whose base case reads 60% but whose downside reads minus 40% is a different proposition from one spanning 20% to 90%, even if the midpoints match; the spread is the risk, and a one-number ROI hides it completely. Weight the scenarios roughly, sanity-check that the downside is survivable at your position size, and only then compare against alternatives. The habit costs five minutes and converts ROI from a persuasion device into an actual decision tool.
When a negative ROI is still the right call
Not every good decision clears the formula, and knowing the exceptions keeps the tool honest. Insurance carries a deliberately negative expected ROI, and buying it is still correct, because the product is not return, it is the removal of ruin. Some spending buys information: a $500 pilot campaign that “loses” money while revealing which channel works has purchased a map, and the map’s value lands in the next campaign’s denominator. And some outlays buy options rather than outcomes, the certification that may never raise your salary but keeps a career door open. The discipline is not to abandon ROI for these, but to name what is actually being bought, price it deliberately, and refuse the lazy version where every unmeasurable purchase gets waved through as “strategic.” Measured exceptions are strategy; unmeasured ones are leaks, and the journal is what tells the two apart at year end.
Frequently asked questions
What is the basic ROI formula?
(What you got back minus everything you put in) divided by everything you put in, times 100.
What is a good ROI on stocks?
The long-run diversified benchmark is 7 to 10% annualized; beating it consistently is genuinely rare.
How do I calculate ROI over multiple years?
Compute total ROI, then annualize with (1 + ROI)^(1/years) – 1, or let the calculator do both.
Can ROI be negative?
Of course: get back less than you put in and the formula reports the loss as a negative percentage, information exactly as useful as a gain.
Is ROI the same as profit margin?
No. Margin divides profit by revenue; ROI divides profit by invested cost. A thin-margin business can carry a high ROI on small capital, and vice versa.
Run any deal through the ROI calculator, upgrade multi-year decisions with the NPV calculator, and the rest of the finance tools keep every percentage honest, including the steady kind built by monthly investing.