Our cheapest lead source has the worst numbers downstream. How do I prove it?
Build one table that multiplies the chain through: cost per lead, divided by contact rate, divided by booking rate, divided by completion rate, gives cost per completed job. Then apply margin by job type. A source with the lowest lead price frequently lands last on that final column. Add handling time and callback rate and the gap usually widens.
One table ends the argument
The reason this debate recurs is that each side is holding a different column. Marketing shows cost per lead. Operations shows what happened after. Neither number is wrong and neither settles anything.
Put them in one row per source and multiply through. Cost per completed job = cost per lead / (contact rate x booking rate x completion rate). Every division by a rate below 1 makes the number bigger, and sources differ enormously in those rates. The final column is the only one worth ranking on.
The costs that never reach the table
- Handling labor. A source requiring many dial attempts per booked job consumes CSR payroll that never appears in a marketing report. Count attempts per booked job by source and price it.
- Cancellation and no-show. A booked job that evaporates consumed a dispatch slot. Cancellation rate by source is one of the sharpest quality signals available and almost nobody tracks it.
- Callbacks and warranty work. Jobs that come back cost margin twice. If one source produces price-driven customers pushed toward the cheapest repair, its callback rate will show it.
- Review and reputation effects. Sources that attract mismatched expectations generate a disproportionate share of poor reviews, which has a cost you will pay later in local visibility.
Guard against the counter-argument you will hear
The usual defense of a cheap source is that its jobs are smaller by nature, so of course the rates look different. Sometimes that is true, and the way to check it is to compare within job type rather than in aggregate. If the cheap source loses inside every job type, mix is not the explanation.
The other defense is volume: the source may be inefficient but it fills the schedule. That argument is legitimate only when you have unfilled capacity and the jobs still contribute positive margin after all of the costs above. Once you are capacity constrained, a low-margin source is actively displacing better work.
Make it a standing report, not a one-time analysis
Source economics drift. Vendors change their mix, auctions change, your own handling changes. An analysis that settled the question last spring will be wrong by fall.
The chain table should run continuously off joined marketing, call and job data, with volume thresholds marking rows that are too thin to rank. That is what a marketing intelligence layer produces, and it is why the underlying integration work matters more than any single report built on top of it.
Topics: lead quality · channel comparison · cost per booked job · reporting
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.