Calculating marketing ROI when the sales cycle takes months
With a six-month sales cycle, the marketing costs from the first quarter belong to the closes in the third. Compare both in the same month and you are not measuring imprecisely but wrongly — and regularly deciding against what works.
In short
- Costs and returns have to be matched to the same case, not the same month.
- Two approaches are workable: a cohort view by entry month, or a rolling twelve-month figure.
- Two proxies report earlier than ROI: the number of qualified enquiries and the time from enquiry to first conversation.
- Your own working time belongs in the numerator. Without it, any comparison between channels is worthless.
Why monthly figures are wrong
An example makes it plain. A company with a six-month sales cycle doubles its marketing budget in January. In February and March costs rise, closes stay flat — ROI looks bad. In April the budget gets cut. In July the closes from the January enquiries arrive while costs are low again — ROI looks excellent.
Both numbers are wrong, and both lead to the wrong decision: cut when it is working, feel vindicated when nothing more is coming.
Two workable approaches
1. Cohort view — more accurate
All enquiries from one month are treated as a group and followed across the full cycle length: how many became customers, with what revenue? Set against that are the marketing costs of exactly that month.
Advantage: matches costs and returns correctly. Disadvantage: the result for January only exists in July.
2. Rolling twelve-month figure — simpler
Costs and returns of the last twelve months are set in relation to each other, updated monthly.
Advantage: available monthly, smooths fluctuations. Disadvantage: reacts slowly to changes and does not attribute individual measures.
For most small companies the rolling figure is the right choice for ongoing reporting, supplemented by a cohort view once a year.
Worth knowing
Cycle length should be stated as a median, not an average. A single deal that took fourteen months shifts the average substantially — and with it the entire matching.
In practice that means: take every closed case from the last year, sort by duration, and take the middle value. That one number determines the period over which costs and returns get combined — making it the most important quantity in the whole calculation.
What belongs in the costs
| Item | Included? | Note |
|---|---|---|
| Media budget | yes | in the month it was spent |
| External providers | yes | in the month the work was done |
| Your own working time | yes | hours times internal rate |
| Tools and licences | yes | apportioned per month |
| One-off build-up work | spread | across its useful life, not one month |
| Sales time after handover | no | that belongs to sales, not marketing |
The third row decides the informativeness. Without your own working time, every channel with no media budget looks free — and the ranking between channels regularly reverses once it is counted.
Two quantities that report earlier
With long cycles, ROI is a look backwards. To know sooner whether a change is working, you need quantities that react faster.
- Qualified enquiries per month. Reacts within weeks and is the most direct consequence of marketing work. Precondition: "qualified" is defined in writing.
- Time from enquiry to first conversation. Reacts immediately and later affects the close rate. It is not a marketing metric in the narrow sense, but it influences ROI more than most marketing measures do.
A pattern that repeats: a company with a nine-month cycle assesses marketing quarterly and keeps concluding that nothing pays off. After three quarters the budget gets cut — exactly when the first cases from the build-up phase are ripening.
The remedy is unspectacular: fix an assessment date matching the cycle length, and until then look only at the inflow quantities. With a nine-month cycle, the first honest ROI statement is possible after around fifteen months — and that belongs said before the start, not after.
What honestly stays unattributable
- Referrals. Someone arriving through a conversation may have read a piece a year ago. That chain cannot be reconstructed.
- Mentions in AI answers without a click. They leave no trace and are probably the larger part of that effect.
- The effect of not acting. Nobody knows what would have happened without the measure.
The only sound proxy is the question in the first conversation: "How did you come across us?" Imprecise, but it captures exactly the cases no measurement reaches.
Help me calculate marketing ROI properly given our long sales cycle. Our figures: - Median time from enquiry to close: [details] - Marketing costs per month, last 12 months: [table] - Own marketing hours per month: [details] - Internal hourly rate: [amount] - Qualified enquiries per month, last 12 months: [table] - Closes per month with revenue and the entry month of the original enquiry, where known: [table] - One-off build-up work in the period: [items and amounts] Tasks: 1. Tell me over what period costs and returns have to be combined for us, and justify it from the median duration. 2. Work out the rolling twelve-month figure. Show the calculation and include our own working time. 3. If the entry months are available: additionally produce a cohort view for the three oldest complete months. 4. Spread the one-off build-up work sensibly rather than charging it to one month. Justify the useful life. 5. Name the earliest point at which an honest ROI statement about a measure started today is possible – and which two inflow quantities we should watch until then. Do not invent figures. Where data is missing, say which.
In closing
With long cycles the monthly comparison is not imprecise but systematically misleading — and it works against the right decision. The remedy has two steps: determine the median cycle length and combine costs and returns over that period.
Until the first result exists, steer by the inflow quantities. And say before the start when the first honest statement is possible — otherwise the budget gets cut shortly before it works.
Common questions
How do you calculate marketing ROI with long sales cycles?
By matching costs and returns to the same case rather than the same month. Two approaches work: a cohort view following all enquiries from one month across the full cycle and setting them against that month's costs, or a rolling twelve-month figure.
Why are monthly figures wrong with long cycles?
Because costs fall in one period and closes in another. Increase the budget and you see only the costs at first and cut — then see good figures, because the closes from the expensive phase arrive while costs are already low. The error is systematic and points against the right decision.
Which cycle length do you use?
The median of all closed cases from the last year, not the average. A single very long case shifts the average substantially and with it the whole matching. That one number determines the period over which costs and returns get combined.
Which metrics report earlier?
Two: the number of qualified enquiries per month, which reacts to marketing work within weeks, and the time from enquiry to first conversation, which reacts immediately and strongly affects the later close rate. Both serve as steering quantities until ROI becomes informative.
Does your own working time belong in the calculation?
Yes, necessarily. Without it, every channel with no media budget appears free, and the ranking between channels regularly reverses once it is counted. What does not belong is sales time after the handover — that counts to sales.
Marketing that sets itself up
The Studio Engine beta is live. Claim your spot and help shape it from the start.
Join the beta →