An application is not a loan, and the fallout is not evenly spread.

Between application and funding sits credit, income, appraisal and the borrower's own change of mind, and a meaningful share never completes.

That fallout is not random across sources: a campaign built on the lowest advertised rate attracts borrowers who will shop again at the first opportunity.

Ranking on applications therefore rewards precisely the campaigns that produce the most fallout, and the cost of that fallout is borne in processing and underwriting time.

Matching funded loans back to campaigns gives the only number that pays anybody, and it routinely reorders a ranking built on application volume.

Keeping unfunded applications in the file is what makes the cost visible rather than merely the outcome.

Refinance volume responds to rates and arrives quickly; purchase volume responds to the housing market and arrives slowly through agent relationships.

Blended, a ranking is dominated by whichever of those conditions prevailed, and it will reverse when conditions do — which reads as marketing failure and is not.

Splitting them gives two rankings that remain meaningful across a rate cycle, and each can carry its own window.

Most lenders find the campaigns that work for one are close to useless for the other, which argues for separate budgets rather than one.

It also stops a refinance boom from being read as evidence that a campaign works, which is the most common error in the category.

Leads are routed to individuals, and conversion varies by who received them at least as much as by where they came from.

A campaign assigned to an officer at capacity will underperform one assigned to somebody with time, and the campaign will be blamed.

Carrying the officer through lets the report hold one constant while examining the other, which is the only way to separate a sourcing problem from a staffing one.

It also identifies where a campaign works for one officer and not another, which is a coaching finding rather than a marketing one.

That distinction is worth protecting: cutting a campaign to fix a routing problem removes the revenue and keeps the problem.

Estate agents, builders and financial advisers send a large share of purchase volume, and that flow is managed by loan officers rather than by marketing.

Because it produces no campaign record, it sits outside the marketing report while carrying much of the business.

Recording the referring partner on the lead puts it into the same comparison on the same funded volume.

The result usually shows a small number of partners producing most of the referral volume, which is a relationship-management decision the report can inform.

It also shows which partners have quietly stopped sending business, which is the half nobody notices because nothing arrives to notice. A partner who sent nine loans last year and one this year is a conversation worth having, and only the funded file will surface it.