The metric is easy to game, so it gets gamed
Nobody sets out to buy bad leads. It happens because cost per lead is the number in the weekly report and it is the easiest number in the funnel to move.
Every lever that lowers it also lowers intent. Wider audiences, vaguer promises, prizes attached to a form. The cost falls and the conversion falls faster.
Count what happens after the form
The useful figures sit downstream: how many leads were reachable at all, how many turned into a real conversation, how many bought.
Once those are visible, expensive channels often turn out to be the cheap ones, because they deliver people who answer the phone.
- Share of leads that can actually be contacted
- Share that reach a real conversation
- Share that buy, by channel and by campaign
- How long each step takes, so slow channels are visible
The hidden cost is your own team
Bad leads are not free. They consume the most expensive hours you have, and they do it invisibly because nobody logs the time spent on people who were never going to buy.
A channel that halves the lead price and triples the volume of unqualified contacts has raised your costs while appearing to cut them.
Qualify earlier, on purpose
Adding friction feels wrong when the target is volume and obviously right when the target is customers. A form that asks one real question filters more than any targeting setting.
The point is not to make the form hard. It is to make it specific enough that people with no intention of buying do not bother.
The arithmetic that settles the argument
Take two channels. One delivers leads at half the price of the other, and one in fifty becomes a customer. The expensive channel converts one in ten.
The expensive channel wins by a wide margin per customer, and the gap grows once you count the hours spent on the forty nine conversations that went nowhere. That time is real, it is paid, and it never appears in the report where the cheap channel looks good.
What to do with the leads you already bought
Stopping a bad channel does not recover what was spent. Sorting the backlog does, at least partly.
Rank what you already have by whatever signals intent, work the top of that list properly, and stop pretending the rest will convert with more follow up. Chasing the bottom of a bad list is the second payment on the same mistake.
Say what a lead means before you buy any
Most disagreements between marketing and sales are definitional. One counts a submitted form, the other counts a person worth calling twice.
Write down which is which before the next campaign. It costs one conversation and it settles arguments that otherwise repeat every month.