If my cost per booked job is acceptable, why does adding budget make it worse?
Because the number you are looking at is an average, and new budget buys at the margin. Marginal cost per booked job is the change in spend divided by the change in booked jobs between two spend levels. It is almost always higher than the average, and it rises as you scale. The average tells you what you paid. The marginal number tells you what the next job will cost.
The arithmetic nobody runs
Average cost per booked job is total spend divided by booked jobs. Marginal cost per booked job is (spend at the higher level minus spend at the lower level) divided by (booked jobs at the higher level minus booked jobs at the lower level).
Run that on your own step-ups and the result is usually uncomfortable. A channel can carry a perfectly healthy average while the last increment of budget you added produced jobs at several times that cost. The average absorbs it and hides it.
Why the marginal curve bends upward
Three forces push it up, and they compound.
What causes it
- Auction pressure. To win more impressions you bid higher on the same inventory, so every click gets more expensive, including the ones you were already winning.
- Intent dilution. Extra volume comes from broader queries, wider radius, or looser audiences. Those leads book at a lower rate, so cost per lead may look stable while cost per booked job climbs.
- Capacity friction. More leads hitting the same phone team means more calls that ring out or sit on hold. The media bought a lead; the operation did not convert it. Call intelligence is how you separate this cause from the other two.
Measuring it without fooling yourself
Change one thing. Hold creative, landing pages and offers constant, step the budget by a meaningful amount, and leave it there long enough to cover both the platform learning period and your own lead-to-job lag. Then compute the marginal number against the prior period, not against last year.
Two traps. Seasonality will happily manufacture a flattering or damning result if you step budget up as demand naturally rises, so compare against the same season if you can. And in-flight jobs mean the recent period always looks worse than it will finally settle, which understates the marginal return of the increment you just made.
When a bad marginal number is still worth paying
A rising marginal cost is not automatically a stop signal. If you have unfilled capacity, an expensive incremental job that still contributes margin beats an idle truck. If you are defending share in a market where a competitor is expanding, the marginal job is buying something the spreadsheet does not price.
What you cannot do is make that call without the number in front of you. Getting spend, leads and completed jobs into one place so the arithmetic is available on demand is the job of marketing intelligence and the paid media reporting sitting on top of it.
Topics: marginal cost · budget scaling · cost per booked job · diagnostics
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.