Why is cost per lead a bad way to judge marketing?
Because a lead is not a unit of value and its definition changes from channel to channel. A cheap lead that books at a third the rate and produces a smaller ticket costs far more per unit of revenue than an expensive one. Cost per lead tells you what you paid to make a phone ring. It says nothing about whether it was a phone call worth having.
The arithmetic that flips the ranking
Suppose source A produces leads at half the cost of source B, and A's leads book at one third the rate. A's cost per booked job is then one and a half times B's, despite looking twice as efficient in the ad report. Add a lower average ticket and the gap widens further.
This is not an edge case. It is the normal relationship between broad, cheap demand and narrow, expensive demand. The cheaper the lead, the more likely it came from someone earlier in their decision, further away, or not qualified at all.
Every channel counts leads differently
A form fill, a chat session, a thirty-second call and a lead purchased from an aggregator are all one lead in a blended report. They are not comparable events. One channel may count a call at fifteen seconds, another at sixty; one deduplicates repeat callers, another does not.
That is why the same month can produce four different lead totals across four systems, and why marketing reporting has to declare a single definition and map everything else to it before any per-lead metric is meaningful.
What to use instead
- Cost per booked job. Same spend, denominator taken from your operational system rather than the ad platform.
- Revenue per lead by source. Blunt, ungameable, and immediately exposes sources whose volume does not convert.
- Gross profit per marketing dollar. The only version that accounts for job mix, and the closest thing to a single number an owner should watch.
When cost per lead is still worth reporting
It is a fine leading indicator. It moves fast, it reacts to auction pressure and creative changes within days, and it will tell you something broke long before booked-job data matures. Treat it as a diagnostic gauge, not a scoreboard.
The failure is using it for allocation decisions. Budget follows booked revenue. If your reporting cannot show booked revenue by source, that is the gap to close first — usually by connecting the ad platforms to the field service system through an integration layer rather than by adding another dashboard.
One more caution. Cost per lead falls automatically whenever a channel starts producing more junk, because junk is cheap and still counts as a lead. A sudden improvement in cost per lead with no change in booked jobs is a warning sign, not a win.
Topics: cost per lead · lead quality · metrics · channel comparison
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.