Every contractor I've worked with tracks cost per lead. It's easy to measure, every vendor reports it, and it moves in the direction you want when you negotiate. The trouble is that nobody signs a lead. They sign a contract, and the path from one to the other is where most of the money goes.
Why cost per lead misleads
Take two sources. The first costs $100 a lead, and one in twenty becomes a sale. The second costs $400 a lead, and one in four becomes a sale. On a cost-per-lead report, the first looks four times better. On a cost-per-sale report, the first costs $2,000 per sale and the second costs $1,600. The "expensive" source is the cheaper way to grow.
Those numbers are an illustration, but the pattern isn't. Purchased leads are usually cheap to buy and expensive to close. They're often shared with other contractors, they skew toward people collecting quotes, and they cost your call center and your reps the same time as a homeowner who searched for you by name.
None of that shows up until you follow each lead all the way to a signed contract.
What it looked like with real sources
As CMO of a PE-backed roofing platform, I traced every lead source to signed contract value. Search and website leads reached approved work far more often than leads bought from third parties, which were closing at roughly a fifth the rate of website leads.
So I cut third-party spend and brought monthly lead count down by about 30%, on purpose. On a lead-count dashboard that looks like a decline. I defended it upward as a quality shift, and the sales math supported it, because the business grows on signed contracts, not on names in the CRM.
The counting rules that make cost per sale work
Cost per sale sounds simple. It stays simple only if a few rules are set before anyone starts reading it.
Measure approved work, not "closed"
"Closed" is a lagging number. A lead that came in last month may not close for weeks, so any recent month looks weak and any older month looks strong. Signed contract value by lead source, read with a standing note on how long leads take to mature, gives a fair read on a young month.
One source, one name
If the same channel appears in the CRM as "Google", "Website" and "Inbound call", its cost per sale is split three ways and none of them is right. Consolidating dozens of source names into a handful of parent sources with accurate sub-sources comes before any of this math.
Keep brand out of channel math
A sponsorship or a brand campaign is real spend, but it isn't a lead source. Counted inside a channel, one large commitment can make that channel look like it doesn't pay. Working spend (media, leads and audiences) gets judged on sales. Fees, tools and brand get tracked separately.
Created and assisted, side by side
Marketing creates some sales outright and helps others that a rep or referral started. Both matter. Report them next to each other and never add them together, or marketing ends up credited twice for the same contract.
Want to see your own sources ranked by cost per sale? Bring your last two quarters to a working conversation.
Book a working conversation →
Thirty minutes. You'll leave with a straight read on where to look first.
What changes when you run on cost per sale
Three things change, usually within a planning cycle.
- The mix changes. Money moves toward sources that produce contracts, even when they cost more per lead, and away from the ones that only fill the calendar.
- The sales floor trusts marketing again. Reps stop chasing leads that were never going to sign, and marketing is measured in the same currency they are.
- The board conversation gets easier. "We spent less on leads and signed more work" is a sentence a PE-backed CEO can take into any room.
One caution. When money moves toward higher-cost sources, lead count usually drops before signed contracts rise. Tell the owner or the board it's coming, show them the cost-per-sale ranking behind it, and set a date to review it together. A planned dip reads as a decision. An unexplained one reads as a problem.
Where to start
Pull every lead from the last two full quarters, with its source, and match each one to its current job status and contract value. Clean up the source names first. Then divide working spend by approved contracts, source by source. The ranking you get will not match your cost-per-lead ranking, and the difference is the budget you can move.
If the data won't cooperate, that's the finding. It usually means the fog is in the lead sources or the CRM, and that's where the work starts.