How to Measure Marketing ROI (Without Fooling Yourself)

The marketing ROI formula, and where the maths goes wrong

The formula itself is genuinely simple:

• Marketing ROI = (Gross profit attributable to marketing − Marketing cost) ÷ Marketing cost
• Multiply by 100 for a percentage, or leave it as a decimal for a multiple (a result of 3 means £3 back for every £1 spent, on top of getting that £1 back).

Two silent substitutions are what break this formula in practice, and both happen quietly enough that nobody notices until the number gets challenged.

The first is using revenue instead of profit. Revenue is everything that came in. It is not what the business actually kept. A campaign that drives £50,000 in revenue on a product with a 20% margin returned £10,000 in profit, not £50,000. Running the ROI formula on revenue instead of margin inflates the result every single time, and the inflation gets worse the thinner the margin is. A high-margin business can get away with this shortcut and still land close to the truth. A low-margin one cannot.

The second is using ad spend as a stand-in for total marketing cost. Ad spend is the easiest number to pull, because it sits in one dashboard with a clean total attached. It is also, on its own, an incomplete picture of what the marketing actually cost. The next section covers exactly what else belongs on that side of the equation.

Here is what the difference looks like with round, illustrative numbers, not a real client's figures. Say a campaign generates £40,000 in revenue on a product with a 25% margin, so £10,000 in profit. If the only cost counted is £4,000 in ad spend, the calculation reads (£10,000 − £4,000) ÷ £4,000 = 150% ROI. Once agency fees, software, and a fair share of staff time are added and the true marketing cost comes to £7,500, the same campaign reads (£10,000 − £7,500) ÷ £7,500 = 33% ROI. Same campaign, same result in the real world, a completely different verdict depending on what got counted.

What actually belongs on the cost side

A defensible marketing cost figure has five components, and most businesses stop after the first one.

• Ad spend. The media bill itself: what was paid directly to Meta, Google, TikTok, or any other platform for placements.
• Agency or freelancer fees. Whether it is a monthly retainer or a project fee, this is a real cost of running the marketing, not a separate line that sits outside the ROI calculation because it "isn't ad spend."
• Software and tools. Analytics platforms, email tools, landing page builders, design software. Small individually, real in total, and almost always left off.
• Creative production. Photography, video, design, and copywriting, whether produced in-house or bought in, is a cost of the campaign that produced the revenue being measured.
• Staff time. The hardest one to count and the easiest one to skip, because nobody wants to work out an hourly rate for the person who spent Tuesday afternoon on the campaign brief. Skipping it doesn't make the time free. It just makes the ROI number wrong.

"The labour time is the one almost everyone leaves out, and it's usually the biggest single gap in the number," Alessandro Lombardo says. "Nine years running The Social Target across 600+ clients, with 50+ active today, and the pattern holds in our experience: the businesses with the most honest ROI number are the ones that stopped pretending their team's time was free." A defensible shortcut, where a full time-and-motion study isn't practical, is to estimate the hours a campaign genuinely took and apply a fully-loaded hourly cost, salary plus the overhead of employing that person, rather than leaving the line at zero.

None of this means every cost needs forensic precision before a number can be trusted. It means the categories need to be present, even as an estimate, rather than silently absent. A rough number that includes all five categories beats a precise number that only includes one.

Why last-click attribution flatters some channels

Attribution is the answer to a question that sounds simple and isn't: when a customer sees three ads and then buys, which ad gets credit for the sale? The model chosen to answer that question changes which channels look like they're earning their keep, sometimes drastically, without a single pound of spend actually moving.

Last-click attribution answers it the simplest possible way. In Google's own words, describing its own last-click model: it "gives all credit for the conversion to the last-clicked ad and corresponding keyword." Whatever the customer clicked immediately before buying gets 100% of the credit. Everything that happened earlier in that customer's journey, the video ad that first caught their attention, the social post that built familiarity with the brand, the blog post that answered their question, gets 0%.

That sounds fair until the pattern becomes obvious: the channels most likely to be the last click are the ones that catch demand, not the ones that create it. Branded search and retargeting ads both tend to appear right before a purchase, because by that point the customer has already decided to buy and is either searching the brand name directly or being shown an ad for something they already looked at. Under last-click, those channels look extremely efficient. What actually happened is that an earlier channel, one that gets zero credit under this model, did the harder job of creating the intent in the first place.

Google itself has moved away from last-click as the default for this reason. Its data-driven attribution model instead "distributes credit for the conversion based on your past data for this conversion action," using the account's own data "to calculate the actual contribution of each interaction across the conversion path" rather than handing all the credit to whichever touchpoint happened to be last. Google no longer supports the first-click, linear, time-decay, or position-based models it used to offer either, though last-click itself is still available for anyone who wants to keep using it. A marketing ROI figure built on last-click data can overstate the channels that close and understate the channels that open, and the gap between the two is exactly the part of the picture a single ROI number, calculated the easy way, will never show.

Payback period vs ROI: two different questions

ROI and payback period answer two different questions, and treating them as the same thing is how a genuinely good result gets scaled at the wrong speed.

ROI asks: was the return worth the cost, overall? It's a ratio, and it doesn't care how long it took to get there. A channel that returns £3 for every £1 spent over eighteen months has the same ROI as one that does it in six weeks.

Payback period asks: how long did it take to get the cost back? It's a measure of time, not return, and it's the number that actually determines whether a business can afford to keep spending while it waits. A campaign can have an excellent ROI and still be a cash-flow problem if the payback period is long enough that the business runs out of runway before the return arrives.

This is exactly the trap covered in more depth in why your best month is really a CAC payback problem: a standout month can look like a clear win on ROI alone, while the customers behind it are still months away from actually paying back what it cost to acquire them. Reading ROI without asking about the payback window is how a business ends up scaling a spike that was quietly borrowing from next year's cash rather than compounding into it.

The practical fix is simple: never report a marketing ROI figure without stating the window it covers, and check the payback period separately before deciding whether a good ROI number is safe to scale.

When ROI is the wrong measure altogether

Some of the most valuable marketing a business runs will show a weak or even negative ROI if measured the standard way, and that isn't always a sign it's failing.

Long sales cycles are the clearest case. A B2B business with a six-month sales cycle will show almost no ROI on this month's marketing spend, because the revenue it produces hasn't arrived yet and won't for months. Measuring that spend against this month's revenue answers a question about timing, not effectiveness.

Brand-building spend runs into the same problem from a different angle. The IPA's published summary of Les Binet and Peter Field's research on the topic names "a growing tension that exists between short-term response activity and long-term brand-building," and warns specifically that using "very short-term online metrics as primary performance measures... has dangerous implications for long-term success." Brand-building activity is built to compound slowly. Judged on a standard ROI window built for direct-response activity, it will consistently look like the weaker channel, right up until it's the reason the direct-response channels are still working at all.

Retention-driven businesses are the third case. A single ROI figure on new-customer acquisition spend misses most of the value a repeat-purchase or subscription business actually generates, because that value shows up in the second, third, and tenth purchase, not the first. Why retention beats acquisition for the value ROI often misses covers the deeper case for why the acquisition-only view undersells what a loyal customer base is actually worth.

None of this is an argument against measuring. It's an argument for matching the measurement window, and sometimes the measurement itself, to what the spend is actually built to do.

What to track instead of chasing one ROI number

A single ROI figure is a useful check, not a full scoreboard, and asking one number to carry the whole picture is how businesses end up either over-trusting a flattering result or dismissing a genuinely good channel that hasn't paid back yet.

The fuller approach runs in two steps that this piece deliberately doesn't repeat in full, because both are covered properly elsewhere. Before calculating anything, get clear on picking the one number that matters in the first place: the single metric closest to an actual sale for that specific business, which the ROI calculation should be built around rather than run in isolation. Once that's settled, what should actually show up in the report built around it covers the reporting layer this article stops short of: how the ROI number, the payback period, and the underlying money metric all belong on the same page, not scattered across different dashboards nobody cross-references.

Nine years, 600+ clients, and 50+ still active on retainer sit behind how we run digital marketing end to end: strategy, paid media, content, and reporting built to answer the ROI question honestly from the start, rather than producing a number that only survives until someone asks what's actually in it.

If the ROI figure currently sitting in a dashboard somewhere hasn't been stress-tested against what it actually counted, the maths above is the place to start. Getting the number right is worth more than getting a bigger one.