Imagine your agency sends over the monthly report and the first slide says, in big letters, that the campaign delivered "a 200% ROI". Sounds like a win. But if you ask how they calculated it and the answer is "we divided sales by what we spent", that number is probably inflated, sometimes so much that the campaign actually lost money. That mix-up is the most common one around ROI, so let's start with the basics. ROI (return on investment) is a metric that tells you how much you gained or lost relative to what you put in. It's expressed as a percentage, and it lets you compare investments of different sizes with the same yardstick.
What is ROI and what does it mean?
ROI stands for return on investment. The question it answers is very concrete: for every dollar (or peso) I put into this, how much came back as profit?
What makes ROI useful is that it normalizes. If one campaign left you 500,000 in profit and another left you 2,000,000, the second one looks better at first glance. But if the first cost 250,000 and the second cost 10,000,000, the story flips. ROI puts both on the same scale and helps you decide where the next dollar should go.
It's used in finance, in technology projects, when buying a machine for a factory and, more and more, in marketing, which is where it gets abused the most. In this article I'll stay mostly in that last territory, because that's where calculation errors end up moving entire budgets.
How do you calculate ROI?
The ROI formula is this:
ROI = (profit - investment) / investment x 100
Let's walk through a made-up example with round numbers so the mechanics are clear (figures in Chilean pesos, but the logic works in any currency). An online specialty coffee shop in Santiago spends $2,000,000 on ads for one month, and that campaign generates $6,000,000 in sales.
If you apply the formula using sales, the result is (6,000,000 - 2,000,000) / 2,000,000 = 200%. Looks spectacular. But the shop doesn't keep the $6,000,000, because it has to pay for the coffee, the packaging, the shipping and the payment processor's fee. Let's say its gross margin is 40%, which means $2,400,000 of those sales is left to cover everything else.
With that number, the correct calculation looks like this: (2,400,000 - 2,000,000) / 2,000,000 = 20%. Still positive, but ten times lower than the first version.
And something is still missing. If an agency ran the campaign and charged $500,000 that month, the real investment was $2,500,000. Recalculating: (2,400,000 - 2,500,000) / 2,500,000 = -4%. The same campaign went from "200% ROI" to losing money, without a single ad changing.
That's why, when someone shows you an ROI, the first thing worth asking is which numbers went on top of the fraction and which went on the bottom. Most of the ROI arguments I've seen in board meetings were, deep down, arguments about exactly that.
What does a 50% ROI mean?
It means you got back what you invested plus half of that amount on top. If you put in $1,000,000, the profit attributable to that investment added up to $1,500,000, and after subtracting what you put in you're left with $500,000 clean. A 100% ROI means you doubled your money, 0% means you broke even, and a negative ROI means you lost part of the investment (at -100% you lost all of it).
What is the difference between ROI and ROAS?
This is the mix-up I see most in digital advertising reports, and it makes sense that it happens, because ad platforms report ROAS and often present it as "return".
ROAS (return on ad spend) divides sales by ad spend. In the coffee example it would be 6,000,000 / 2,000,000 = 3, which shows up in reports as "3x ROAS". It's a useful metric for comparing campaigns or ads within the same platform, because it's quick to calculate and doesn't require knowing your margins.
What ROAS doesn't tell you is whether you made money. A ROAS of 3 can be a great deal for a product with a 70% margin and a loss for one with 25%. A practical way to connect both metrics is to calculate your break-even ROAS, which is 1 divided by your gross margin. With a 40% margin, break-even ROAS is 2.5, and any campaign below that line is losing money even if the report looks green. With a 25% margin you need a ROAS of 4 just to break even.
If you're just getting started with paid ads, in what is advertising I explain how the auctions that set your cost per click work, which is the other half of this equation.
Which costs are usually forgotten when calculating ROI?
On the investment side, the four most frequent omissions are:
- Internal team hours. If two people spent half their workday for a month on a campaign, that has a cost even if it never shows up on an invoice.
- Agency and freelancer fees, plus tools (the email software, the SEO tool license, the designer who made the assets).
- Production: photos, videos, landing pages, translations.
- Discounts. If the campaign offered 20% off, that discount is part of what you invested to sell.
On the profit side, the typical error goes the other way: counting sales that would have happened anyway. If you send a coupon to customers who were going to buy from you that week regardless, the campaign takes credit for sales it didn't generate. The rigorous way to measure this is with a control group (you send nothing to part of your base and compare how much each group bought), and most email marketing tools let you do this at no extra cost.
How do you calculate ROI when results take time?
This is where ROI gets more interesting, and also easier to misread. Some investments pay back in days, like a paid search campaign, and others take months, like content, organic search or a well-nurtured customer base.
Let's use SEO as the example, since it's the territory I know best, although the logic applies to any long-term channel. Imagine (again, made-up numbers) a B2B software company that invests $3,000,000 a month in content and optimization. For the first three or four months, organic traffic barely moves and monthly ROI is deeply negative. Around month eight a steady flow of demo requests starts coming in, and since the articles keep bringing visits even if you stop publishing, cost per customer drops month after month. If you evaluate the channel in month three, you shut it down. If you evaluate it at 18 months, it may turn out to be the company's best-returning channel.
Something similar happens, with even more noise, with visibility in ChatGPT answers or Google's AI Overviews. Some people discover your brand there and search for you by name days later, and that path rarely gets recorded as "organic" in your analytics. If you only count what the tool attributes directly, you'll systematically underestimate those channels.
Three adjustments help measure these cases fairly:
- Set the time window before you start. Saying "we'll evaluate this channel at 12 months" keeps someone from cutting it in month two out of anxiety.
- Use customer lifetime value instead of the first sale. If an average customer buys from you four times a year, the profit from acquiring them includes all four purchases. To calculate that you need clean repeat-purchase data, which is exactly what a well-fed CRM gives you.
- Annualize before comparing. A 30% ROI over three years works out to less than 10% per year, so it performs worse than a 20% ROI achieved in twelve months.
What is a good ROI?
The honest answer is that it depends on what you compare it against. A good ROI is one that beats the best alternative you had for that same money. Let's say parking the money in a term deposit would have paid 5% a year (a reference number, rates change all the time). In that scenario, a campaign with a 4% annual ROI left you worse off than doing nothing with the money, even though the number is positive.
Within marketing, comparing channels against each other using the same calculation method is more useful than hunting for a magic percentage. If all your channels are measured with margin, full costs and the same time window, the resulting ranking is reliable even if each individual number carries some error. If each channel is measured its own way (the ad platform with ROAS, email with gross sales, SEO with traffic), comparison becomes impossible and the budget ends up flowing to whichever channel reports on itself most flatteringly. If you're building that dashboard, how to choose the right KPIs and the metrics for measuring organic go into more detail on what to include.
ROI also has limits. Some investments have a real but diffuse effect, like building a brand or getting press coverage, where the return arrives through many doors at once. For those it makes sense to track intermediate indicators, such as searches for your brand name or sales to customers who arrive with no campaign attached, and accept that an ROI accurate to the decimal isn't going to exist.
To wrap up: a five-column spreadsheet
If you want to start measuring your marketing ROI more rigorously, you don't need a new tool. Open a spreadsheet and build one row per channel with five columns: total monthly investment (including team hours, agencies and tools), attributed sales, gross margin on those sales, net profit (margin minus investment) and ROI. Fill it in for the last three months using the same criteria for every channel, and add the break-even ROAS of your main products next to it. Chances are at least one channel will change places compared to what you believed, and with that spreadsheet in hand, your next budget conversation starts from numbers everyone reads the same way.