A UAH 200 lead and a UAH 200 customer are different things
This is the most common substitution in small business. The agency reports cost per lead, the owner hears cost per customer, and both are satisfied until the quarter closes.
Do the arithmetic. The channel delivers a lead for UAH 200. One in ten buys. The customer therefore costs UAH 2,000, not 200. If your average order is UAH 1,800, the channel is not merely unprofitable — every new customer takes money out.
The agency has broken nothing. It is accountable for cost per lead, and that number really is good.
The formula that does not lie
CAC = (ad spend + salaries + contractors + tools) ÷ NEW customers
Four notes on the numerator, each of which people skip.
Salaries count. An in-house marketer's salary is part of acquisition cost. Leave it out and you are comparing a channel against your own free labour.
Contractors count. An agency retainer is acquisition spend, not a separate line.
Tools count. CRM, analytics, call tracking, creative software.
Only new customers in the denominator. A repeat purchase is not acquisition. Mix them in and CAC drops beautifully while the business notices nothing.
Three places the maths breaks
Analytics conversions instead of sales. Google Analytics counts as a conversion whatever you told it to count: a button press, an opened form, a scroll depth. A customer is someone who paid. With no link between advertising and actual payment you are computing an invented number, and no formula fixes that.
Different deal cycles. In a niche with a three-month cycle, September's spend produces December's customers. Divide September spend by September customers and you get another period's CAC. Work in cohorts: what the customers acquired from September traffic cost, whenever they eventually paid.
Mixed channels. Someone sees an ad, later arrives from search, later still types the address directly. Last-click attribution credits the direct visit, and the advertising looks redundant. That is not a reason to buy an attribution platform for thousands a month — it is a reason not to switch a channel off on the strength of one report.
What ROMI is acceptable
Honestly: in isolation from your margin, none.
ROMI is calculated on gross profit, not revenue:
ROMI = (revenue × margin − acquisition spend) ÷ acquisition spend × 100%
Revenue of UAH 85,000 at a 35% margin is UAH 29,750 of gross profit. Spend UAH 20,000 and ROMI is 49% — every unit of spend returns about 1.49 of gross profit. The same UAH 85,000 at a 12% margin yields UAH 10,200, which is a ROMI of minus 49%. Identical revenue, opposite verdict.
So "what ROMI is good" has no answer without the margin. The calculator on the home page works exactly this relationship — put your own three numbers in.
Our position: if you cannot state your margin to within five percentage points, the place to start is management accounting, not advertising. The objection holds where a business sells one homogeneous product at a stable markup — there the margin fits on a napkin, and you can go ahead and count.
An hour of arithmetic
- Take one channel and one month. Not the whole marketing function — you will start and abandon it.
- Add every cost attached to that channel: budget, fees, a share of salaries, tools.
- Pull the customers who paid for the first time in that period and came through that channel.
- Divide. That is your CAC.
- Multiply average order value by margin. If the result is below CAC, the channel is losing money right now.
- Repeat for a second channel. Two honest numbers compared beat the perfect analytics you do not have.
If step three reveals that no link between customer and channel exists, that is itself the finding. The first month goes into building it, and only then does a conversation about budgets mean anything.