How do you know if a marketing channel is working?

Channel decisions are normally made on cost per lead because that is the number every platform reports. It cannot settle the question, and the reason is simple: channels differ enormously in what proportion of their leads become customers. A channel producing leads at half the price of another, converting at a third of the rate, is more expensive per customer and looks cheaper in every report either party produces.

The number that decides it is cost per acquired customer, and computing it requires exactly one thing most businesses do not have: source recorded on the customer, not just on the lead. Total spend on the channel over a period, divided by the customers it produced in that period, including the months where it produced none. That is comparable across channels in a way nothing upstream of it is.

It becomes a decision when held against what a customer is worth. A channel costing less than customer value is worth expanding; one costing more is worth fixing or stopping. That comparison also resolves the argument that recurs whenever a channel looks expensive in isolation — expensive relative to what is the missing half, and the answer is relative to what the customer produces.

Attribution is the part that generates the most argument and deserves the least of it at small-business scale. Customers encounter a business several times before contacting it, so no method assigns credit perfectly, and elaborate models mostly move the imprecision around. Asking the customer how they found you, recorded consistently, is imperfect in known ways and good enough to distinguish a channel that produces customers from one that does not. That distinction is what the decision needs.

The timing rule matters as much as the metric. A channel must be measured over a period at least as long as the business's own enquiry-to-revenue cycle, or the assessment measures activity and calls it a result. Where that cycle is a quarter, a monthly channel review will consistently favour whatever generates the most immediate motion, which is rarely the same as what produces the most customers.

One check before any channel is judged: whether the leads it produced were actually handled. A channel delivering enquiries into an intake that answers slowly or not at all will show a poor cost per customer that has nothing to do with the channel, and businesses regularly cancel a working channel for a failing that would have affected every other channel equally. Response rate and response time on that channel's leads should be looked at in the same session as its cost.