How do I calculate cost per funded loan?

Divide everything a lead source cost you over a period (lead fees plus any labor and tooling that source required) by the number of loans that funded from that source in a matched window. Fund dates lag lead dates by weeks or months, so attribute funded loans back to the month their lead arrived.

Why is cost per lead a misleading metric?

Because a lead is an input, not an outcome. A cheap lead sold to five competitors with no credit policy can cost more per funded loan than a lead at several times the price that you alone receive. Cost per lead compares sticker prices; cost per funded loan compares businesses.

What conversion stages sit between a lead and a funded loan?

Typically: delivered lead → contact → appointment or real conversation → application → funded loan. Each stage has its own rate, and each is influenced by different things: lead quality and exclusivity early, your process and follow-up in the middle, borrower qualification at the end.

How many leads do I need before cost per funded loan is meaningful?

Enough that one loan either way doesn't swing the number wildly, which usually means judging a source over full allocation cycles, not a first handful of leads. Small samples measure luck. Consistent measurement over a quarter tells you something you can act on.

What data do I need to track it?

Two linked records: the vendor's delivery data (which leads arrived, when, from which source, with credits netted out) and your pipeline outcomes (which of those leads applied and funded). A per-lead delivery log plus a CRM that tags lead source covers it.

Cost Per Funded Loan: The Only Lead Metric That Matters

Cost per funded loan is your total spend on a lead source divided by the number of loans that actually funded from it. It's the only lead metric that survives contact with reality: cost per lead can be gamed by a low sticker price, contact rates can be inflated by counting voicemails, but a funded loan either happened or it didn't. If you track one number per lead source, track this one.

Why the obvious metric misleads

Cost per lead is the number every vendor advertises, because it's the number they control. It answers the question "what does an input cost?" And buying inputs is not the business you're in.

The trouble is that a dollar figure per lead smuggles in an assumption: that all leads convert alike. They don't, and the differences aren't small. A lead sold to you alone converts differently from the same lead sold to five buyers racing to dial first. A lead with a wrong number converts at zero, and whether you were credited for it changes your real spend. A lead that arrives scored, with a booking link in the first text, converts differently from a name on a spreadsheet delivered in tomorrow's batch.

None of that is visible in cost per lead. All of it is visible in cost per funded loan.

The arithmetic (with honest numbers)

The formula is division; the discipline is in what you divide.

The numerator is everything the source costs: lead fees, minus any credits actually received, plus the labor the source demands (if a shared-lead program only works with a dialing rotation, that payroll belongs here), plus any tooling you bought specifically to work it.

The denominator is funded loans attributable to the source, matched to the right window. Funding lags lead arrival by weeks or months, so tie each funded loan back to the month its lead arrived. Judging July's spend by July's fundings mostly measures your pipeline's lag, not the source.

To see why the comparison flips, run the arithmetic on two hypothetical sources, using purely illustrative numbers chosen to make the mechanics visible. Source X sells leads at a low unit price; you buy a large batch, but between shared resale, uncredited junk, and slow delivery, only a small share become conversations and one loan funds. Source Y charges several times more per lead for exclusive, scored, fast delivery; you buy far fewer, but a much larger share become appointments and two loans fund. Divide each source's total spend by its fundings and the "expensive" source is routinely the cheaper business: the sticker relationship and the outcome relationship point in opposite directions. Your own numbers will differ; the point is that only this division reveals which way they point.

See exactly how your leads would be delivered

Twenty minutes, live screens, per-market pricing in writing, and a straight answer if your market is full.

Book a call

What actually moves the number

Cost per funded loan decomposes into stages (delivered → contacted → appointment → application → funded), and each stage has different owners.

Contact rate is driven by lead validity, exclusivity, and your speed to lead. Wrong numbers and five-way resale races kill it before skill enters the picture. This is the stage where vendor quality shows up most nakedly.
Appointment rate is where process lives: scoring that tells you what to work first, follow-up that doesn't quit after two attempts, self-booking that catches borrowers you couldn't reach live.
Application and pull-through are increasingly about fit and qualification: your consultation, your product set, the borrower's actual situation.

The practical use of the decomposition: when your cost per funded loan is bad, it tells you where to look. Terrible contact rates point at the vendor (or your response time). Good contact but few appointments points at follow-up. Good appointments that don't fund points at qualification or the consultation itself. One metric on top, a diagnosis underneath.

The measurement plumbing

You can't compute any of this without two datasets agreeing with each other:

Delivery data from the vendor: which leads arrived, when, with invalid credits netted out. This is where a per-lead delivery log earns its keep: it gives you an auditable count of what you actually paid for. (It's why every Mortgage Connect Pro lead carries one, and why our monthly report shows delivered-versus-contracted and credits itemized: the numerator of this calculation, handed to you reconciled.)
Outcome data from your CRM: every lead tagged by source, and its stage transitions recorded through funding.

The join between them is the whole game: lead in, source tagged, outcome recorded. If your CRM can't tell you which funded loans came from which source, fix that before buying another lead from anyone, because you're currently unable to know which of your vendors is good.

The bottom line

Cost per lead tells you what a record costs; cost per funded loan tells you what a customer costs, which is the number your business actually runs on. Compute it per source, over matched windows, with labor and credits included. Let it, not the sticker price, decide where next month's budget goes. Every other lead metric is a stage in this one.