How to Measure Marketing ROI You Can Defend
The formula for ROI is not the hard part. Defending the number it produces is.
Anyone can compute (revenue − cost) ÷ cost. What separates a number that survives a boardroom from one that gets picked apart is whether you can explain, on the spot, exactly what "revenue" meant in that calculation - and whether the same method would have produced a believable number a year ago too.
The ROI formula, and where it actually gets contested
The formula is never the contested part. The contested part is "revenue attributable to marketing" - which requires an attribution model, a time window, and a decision about how to treat customers who would have converted anyway. State those three choices next to the number, every time. A ratio with no stated method is a claim; the same ratio with its method attached is evidence.
ROAS: what "good" actually depends on
Return on ad spend (ad revenue ÷ ad cost) is a narrower, easier metric than full marketing ROI, and it's frequently misread because a single "good ROAS" number gets repeated as if it applies to every business. It doesn't. A 3x ROAS on a product with 70% margin leaves real profit; the same 3x on a product with 15% margin can be a loss once fulfillment and returns are counted.
Hypothesis, not a measured fact about your business: a commonly cited rule of thumb puts a "healthy" ROAS somewhere around 3x-4x for many ecommerce categories. Treat that as a starting benchmark to test against your own margin structure, not as a target to defend on its own.
Customer acquisition cost (CAC)
The defensibility risk here is scope: which costs count. Ad spend is obvious; salaries, tools, agency retainers, and content production often get left out to make CAC look smaller. If someone in finance recalculates CAC with a fuller cost base and gets a different number than yours, the mismatch damages trust in every other metric you present - even the accurate ones.
Building the report itself: for the boss, the CEO, or the board
A report meant to justify a marketing budget or show marketing results to executives should carry three things every dashboard-style export skips: the methodology (attribution model and window), the comparison point (what changed when spend went up, down, or paused), and an honest gap section (where the number disagrees with another system, and why). That last section is the one people trust most, because it is the one dashboards never include.
If your channel numbers already disagree with each other before you even get to ROI, the report will not survive scrutiny no matter how it's formatted - see why analytics numbers don't match across platforms for the underlying cause.
Proving ROI, ROAS, and CAC
▸How do you prove marketing ROI?
You prove it by showing the methodology alongside the number: what counts as attributable revenue, what attribution window and model was used, and what the same calculation would have shown before the marketing activity ran. A number without its method is an assertion, not proof.
▸How do you calculate marketing ROI?
The formula itself is simple: (Revenue attributable to marketing − Marketing cost) ÷ Marketing cost. The work is in defining "revenue attributable to marketing" consistently and defensibly - which requires a documented attribution approach, not just the formula.
▸What is a good ROAS?
There is no universal good ROAS - it depends on your margin, average order value, and customer lifetime value. A 3x ROAS can be excellent on a high-margin product and unprofitable on a low-margin one. Treat any flat benchmark you read (including "3x-4x is good") as a rule of thumb, not a fact about your business - verify it against your own margin structure before acting on it.
▸How do you calculate customer acquisition cost (CAC)?
CAC = total sales and marketing spend over a period ÷ number of new customers acquired in that period. The defensibility question is what counts as "sales and marketing spend" - if you leave out salaries, tools, or agency fees to make the number look better, it will not survive a finance review.
▸How do I show marketing is working to executives who are skeptical?
Show the calculation, not just the conclusion. A one-line "ROI is 4.2x" invites the question "how do you know." A report that shows the attribution method, the time window, and what changed when marketing paused or scaled invites a different, better conversation.
Build a reporting layer that survives the follow-up question.
We engineer the attribution and reporting structure underneath the ROI number, so the methodology is real, not decorative.