Facebook Ads ROI Calculator
Return on ad spend as reported, as profit against the break-even the margin sets, and as incremental profit once the conversions you had anyway are removed.
Ad spend $12,000.00
Agency, tools and fees $1,500.00
Total cost $13,500.00
Revenue the platform claims $41,000.00
As reported
return on ad spend 3.42×
break-even return 2.08×, which is one over the 48% margin
reading 1.64× the break-even return
As profit
gross profit on that revenue $19,680.00
less the cost $13,500.00
profit $6,180.00
profit a unit of spend $0.4578
As incremental profit
share that is incremental 65%
incremental revenue $26,650.00
incremental return 2.22×
incremental profit -$708.00
verdict profitable as reported and not profitable once the conversions that would have happened anyway are removed
Against the whole business
total revenue $96,000.00
blended return 7.11×, all revenue over all marketing cost
attributed share 42.7% of revenue is claimed by this channel
worth watching the blended figure, because it cannot be inflated by attribution
What the return would have to be
at a 20% margin 5.00× to break even
at a 35% margin 2.86× to break even
at a 50% margin 2.00× to break even
at a 65% margin 1.54× to break even
How much can be wrong before it stops working
incremental share at break-even 68.6%
which means anything above that share being real keeps the campaign profitable, and anything below it does not
A 3.42× return on a 48% margin is measured against a break-even of
2.08×. Return on ad spend quoted without the margin is a number that
cannot be judged: the same 3× is comfortable at 60 percent and a loss at
25.
At 65% incremental, the campaign loses $708.00. Break-even sits at 68.6%
of the attributed conversions being real, which is the number a holdout
test would give you and the only way to get it.
A platform reports conversions it observed within its attribution
window, not conversions it caused. Someone who was going to buy, saw an
ad on the way and bought, is counted in full, and retargeting is the
worst offender because it advertises to people already in the funnel.
Attribution settings change the answer without changing the business. A
7-day click and 1-day view window reports more than a 1-day click window
on the same campaign, and comparing two channels on different windows
compares reporting conventions.
Blended return, all revenue over all marketing cost, is the figure that
cannot be inflated by attribution. It is blunt and it moves slowly, and
it is the right check on a channel report that keeps improving while the
bank balance does not.
Gross margin here excludes fulfilment, payment fees and returns. If
those are material, use contribution margin instead: the break-even
return rises and several campaigns that looked fine stop looking fine.
First-order return is not the whole return either. A first purchase that
becomes a repeat customer earns more than this arithmetic credits it
with, which is the honest argument for accepting a lower return on
acquisition, as long as the repeat rate is measured rather than assumed.
Output is valid and updates as you type.
Fix the highlighted fields to update the output.
A 3.42× return on ad spend sounds like a result. Two corrections sit between it and profit, and on the example they turn it into a loss of $708.
The first is margin. Break-even return is one divided by your gross margin, so at 48 percent it is 2.08×. The same 3× return is comfortable at 60 percent margin and a loss at 25, and the figure quoted without the margin cannot be judged at all.
The second is incrementality. A platform counts conversions it observed, not conversions it caused, and a share of what it observed would have happened anyway: brand searches, returning customers, people already on their way to buying. Holdout tests at large advertisers have repeatedly found a meaningful part of claimed conversions is not incremental, and the only way to learn your share is to run one.
How to use
- Put in the ad spend, the revenue the platform claims, and your gross margin.
- Add agency fees and tools. They are part of the cost of the channel.
- Set the incremental share from a holdout test if you have one. If you have not, leave it at 100 and read the result as the platform’s own account of itself.
Example
$12,000 of spend and $1,500 of fees against $41,000 of claimed revenue at a 48 percent margin, with a holdout suggesting 65 percent is incremental:
As reported
return on ad spend 3.42×
break-even return 2.08×, which is one over the 48% margin
reading 1.64× the break-even return
As profit
gross profit on that revenue $19,680.00
less the cost $13,500.00
profit $6,180.00
As incremental profit
share that is incremental 65%
incremental revenue $26,650.00
incremental return 2.22×
incremental profit -$708.00
verdict profitable as reported and not profitable once the conversions that would have happened anyway are removed
Against the whole business
total revenue $96,000.00
blended return 7.11×
attributed share 42.7% of revenue is claimed by this channel
How much can be wrong before it stops working
incremental share at break-even 68.6%That last line is the one to carry into the meeting. The campaign works if more than 68.6 percent of the claimed conversions are real, and nothing in the platform’s reporting can tell you whether they are.
Pitfalls
Return on ad spend without a margin is not a number. Break-even is one over the margin. Quote them together or not at all.
Retargeting is where incrementality is worst. It advertises to people who are already in the funnel, so it takes credit for conversions that were arriving anyway. It usually reports the best return in the account and it is the first thing a holdout test deflates.
Attribution windows change the answer without changing the business. A 7-day click and 1-day view window reports more than a 1-day click window on the same campaign. Comparing two channels on different windows compares reporting conventions.
Agency fees and tools are part of the cost. A 15 percent management fee moves the break-even return by 15 percent, and it is left out of most versions of this calculation.
The blended figure is the honest check. All revenue over all marketing cost cannot be inflated by attribution. It is blunt and slow, and it is the right thing to watch when the channel reports keep improving while the bank balance does not.
Gross margin is not contribution margin. Fulfilment, payment fees and returns come out before you have earned anything. Where they are material, use contribution margin and expect the break-even return to rise.
First-order return is not the whole return. A first purchase that becomes a repeat customer earns more than this credits it with, which is the honest case for accepting a lower acquisition return, as long as the repeat rate is measured.
Compatibility
Arithmetic in the browser: nothing is uploaded and nothing is stored.
The three views use the same inputs so they cannot disagree: reported return divides revenue by spend, profit applies the margin and subtracts every cost including fees, and the incremental view scales the revenue before doing the same. The test suite asserts the sign changes between the second and third on the example, since that reversal is the reason the tool exists.
Break-even return is one over the margin, printed next to the reported figure rather than in a separate section, and the sensitivity table shows it at four margins so the effect of the margin on the verdict is visible without re-running anything.
The incremental share at break-even is total cost divided by gross profit on the claimed revenue, which is the threshold a holdout test result would be compared against.