Is a marketing agency worth it? How to answer it with numbers

Aug 18
3 min read

Updated: 4 days ago

Introduction

This question usually gets answered with opinions, which is why it stays unresolved. It is actually arithmetic, and the arithmetic is not complicated.

The catch is that you need two numbers first, and most businesses asking the question do not have them.

1. Is a marketing agency worth it? Run this calculation

You need the gross profit an average customer produces over their whole relationship with you, and the total cost of the engagement.

Divide the cost by the profit per customer. That gives you the number of additional customers required to break even.

If a customer produces $300 in gross profit and an engagement costs $12,000, you need 40 additional customers to break even. Now the question becomes answerable: is 40 additional customers plausible from this work, in the period you care about?

That is a question you can reason about. "Is it worth it?" is not.

2. If you cannot run the calculation, that is the finding

Most businesses cannot answer it, because they do not know their customer lifetime value or their acquisition cost.

If that is you, the honest answer is that you cannot yet evaluate any marketing purchase — and the first thing worth buying is the measurement, not the campaigns. It is the cheapest piece of work available and it makes every later decision informed.

Buying execution before you can measure it is how businesses spend for years without knowing what worked.

3. When the answer is clearly no

Some situations make the answer straightforward.

You have not established that anyone wants what you sell. Marketing amplifies demand; it does not create product-market fit. Amplifying something nobody wants is expensive.

Your margins cannot support acquisition costs. If a customer produces very little gross profit, paid acquisition of any kind may be structurally unaffordable, and no agency changes that.

You cannot handle more customers. Booked out with a waiting list means the constraint is capacity or price, not demand.

You want someone to take it off your hands entirely. Achievable, but it produces dependency rather than capability, and you should choose it knowingly.

4. When the answer is usually yes

You know demand exists and cannot capture it. Enquiries arrive and leak, or you are visible to the wrong people. That is a structural problem with a known shape.

You have a specific problem you cannot diagnose. Customers arrive but do not return, or profit is flat while revenue grows. Diagnosis is genuinely specialist.

The expertise gap is large and the work is finite. Setting up measurement properly, structuring accounts, working out pricing — these take years to learn and weeks to buy.

5. Worth it and worth continuing are different questions

An engagement can be worth starting and not worth extending.

The specialist work — diagnosis, structure, measurement, pricing — has an end. What remains afterwards is operational, and operational work is usually cheaper done internally from a written procedure.

Judge the extension on its own arithmetic rather than on how the first phase went.

6. What actually determines the answer

Not the agency's talent. Two things: whether you are buying the right work, and whether anything remains afterwards.

Ask what will exist at the end that does not exist now. Accounts in your name, documented processes, a dashboard you can read, a trained person. If the answer is nothing, the cost has to be justified purely by results during the engagement — a much higher bar.

7. Reduce the cost of being wrong

Whatever you conclude, structure the first commitment so a mistake is survivable: a defined scope, staged payments, and a clear deliverable.

A small project tells you how someone communicates, whether they disagree with you when they should, and whether the work is good — none of which is visible beforehand. Expand on evidence.

Conclusion

Divide the cost by the gross profit per customer to get the break-even number of customers, then judge whether that is plausible. If you cannot run that calculation, buy the measurement first.

The answer is no when demand is unproven, margins cannot carry acquisition, or you are already at capacity. It is usually yes when demand exists but leaks, or when you need diagnosis. Either way, require that something remains when it ends.