Marketing payback: how to calculate the return on spend
The question «does this pay off» decides the fate of any channel, but most calculations break on two things: the work is left out of the costs, and revenue is used instead of profit. Here is how to calculate it so that the number means what you think it means.
Calculate on margin, not revenue: subtract cost of delivery from income, then divide by all the channel's costs — budget plus contractor's work plus staff time. A figure above one means the channel earns. The most common mistake is calculating on turnover and forgetting the work: that way almost any channel looks profitable.
In this article8 sections
What exactly we are calculating
There are two similar metrics, and they get confused constantly. The difference is fundamental.
| Metric | How it is calculated | What it shows |
|---|---|---|
| Return on turnover | Revenue ÷ channel costs | How much turnover each unit spent produces |
| Return on margin | (Revenue − cost of delivery) ÷ costs | How much each unit spent earns |
The first almost always looks excellent and means almost nothing: at a 20% markup, turnover three times your costs is a loss. Management decisions are made on the second.
A guide for the second: below one the channel is losing money, around one it breaks even, from one and a half upwards it earns. The exact threshold depends on your markup.
What goes into costs
This is where most calculations break. The ad budget is only part of a channel's cost.
- The ad budget paid to the platform.
- The contractor's fee or the specialist's salary.
- Creating materials: texts, photographs, landing pages.
- Staff time spent handling enquiries.
- Services — analytics, dynamic numbers, CRM — in the share attributable to the channel.
The last two are the ones most often skipped. If a manager spends two hours a day on enquiries from one channel, that is part of its cost.
What goes into income
Not the total value of deals but what is left after the direct costs of delivering them: materials, contractors' work, commissions.
- Collect the deals for the period by channel.
- Subtract the cost of delivery from the total.
- Count from the date of first enquiry, not the date of payment, if the cycle is long.
- Attribute a client's repeat purchases to the channel that first brought them.
A worked example
A small company, one month, paid search.
| Line | Amount |
|---|---|
| Ad budget | $800 |
| Specialist's work | $300 |
| Landing page (this month) | $100 |
| Total costs | $1200 |
| Enquiries | 60 |
| Clients | 12 |
| Revenue | $6000 |
| Cost of delivery | $3600 |
| Margin | $2400 |
| Return on margin | 2.0 |
The channel earns: every dollar spent brought two dollars of margin. Calculated on turnover it would read 5.0 — a prettier and less useful figure.
A long sales cycle
If months pass between enquiry and payment, a monthly calculation lies: the costs fall in this month, the income in a future one.
- Work by cohorts: match a month's costs against income from that month's enquiries, even when it arrives later.
- Read the data with a lag equal to your cycle.
- Recent months always look worse — that is normal, not a failure.
- For such niches an annual calculation is more accurate than a monthly one.
Should repeat sales be counted
If clients come back, a one-off calculation understates the return. Then you count not the first deal but what a client brings over their lifetime.
A practical approach for a small business: take the average number of purchases a client makes in a year and the average margin per purchase. The product is what a client actually brings. That is the figure to compare acquisition cost against.
It is precisely repeat sales that can make a channel profitable when a one-off calculation shows it losing. But you can only use that argument if the repeats genuinely exist and have been counted.
Common mistakes
- Calculating on turnover. Gives an inflated and pleasant figure.
- Forgetting the work. Counting only the budget is the most widespread error.
- Crediting every sale to the last channel. Devalues whatever brings people in first.
- Using too short a period. One large deal distorts a month.
- Ignoring seasonality. Comparing December with February is pointless.
- Not accounting for refusals and refunds. Paid and not cancelled are different sums.
What to do with the result
| What you got | Likely cause | Action |
|---|---|---|
| Below 1 | Expensive acquisition or thin margin | Narrow the queries, raise the price, check handling |
| Around 1 | The channel breaks even | Improve conversion rather than raise the budget |
| 1.5–3 | A healthy state | Increase spend gradually |
| Above 4 | Most likely costs were undercounted | Recheck that everything is included |
The last row is not a joke: a result that is too good almost always means part of the costs never made it into the calculation.
Frequent questions
What return counts as normal?
It depends on your markup. At around 40% margin, a return on margin of 1.5–2 is healthy. Using someone else's figures without accounting for your own economics is pointless.
How do we calculate when sales happen offline?
Ask people where they heard about you and log the answer alongside the deal value. Accuracy is lower, but the overall picture by channel will be right.
Should a manager's salary be included?
Yes, in proportion to the time spent on that channel's enquiries. Otherwise a channel with cheap but labour-intensive requests looks better than it is.
What period should we use?
A month is the minimum for operational decisions, a quarter for strategic ones. With a long sales cycle only a quarter or a year gives a reliable picture.
What if channels overlap?
Look at both extremes: first touch and last touch. If a channel looks bad in one and good in the other, it works at the top of the funnel and must not be switched off.
Should organic traffic be included?
Yes, but its cost is the cost of the promotion work, not zero. Otherwise organic always looks infinitely profitable.
The materials answer general questions. We will look at your specific case — free and without obligation.