Mortgage Lead Scoring: Triage Before the First Call

Every mortgage lead list hides three different populations: files ready to move this month, files that will be ready someday, and inquiries that will never become files at all. A brokerage that calls them in arrival order treats all three identically, which means the ready-now borrower waits behind the curious browser: and in a trade where the first substantive callback usually wins the file, that ordering error is quietly expensive. Lead scoring is the fix: a triage layer that ranks every lead BEFORE the first call, so broker hours land where funded deals live. Here is how to build it, drawn from the scored-and-triaged architecture Nova built for Niche Mortgages.

Why mortgage leads specifically need triage

• The funnel is long and the mix is wide: most leads are months from transacting (research-stage borrowers doing scenario math), while a minority have an accepted offer and a financing condition expiring in nine days. Same inbox, wildly different clocks.
• Broker time is the scarce resource: one funded file is worth a four-figure commission, so an hour spent qualifying a never-borrower is an hour taken from a file that would have funded. Cost per funded deal (the metric that matters) is mostly a time-allocation problem wearing a marketing costume.
• Speed to lead only works with a target: 'call everyone within five minutes' is impossible at volume; 'call the hot lane within five minutes' is a system. Triage is what makes the speed rule executable.

What to score on (the five signals)

• Timeline: the king signal. 'Accepted offer' or 'renewing within 120 days' outranks everything; 'just exploring' routes to nurture regardless of how strong the file is.
• Purpose: purchase with offer, pre-approval hunt, renewal/switch, refinance, decline rescue: each lane has its own urgency curve and its own call script.
• File strength: income shape, self-assessed credit, down payment or equity story: not to reject weak files (hard files are a specialty, not a disqualification) but to route them to the broker who works them best.
• Engagement depth: a visitor who completed a nine-question funnel and read three articles is not the same lead as a name scraped from a contest form, even with identical answers.
• Source: tag every lead's origin (funnel, calculator, referral, campaign, seminar) so scores can flow BACK to the channels: which is how marketing budgets learn.

The architecture lesson: capture should do the scoring

The cheapest place to score a lead is the moment it's created, by the questions the capture asks. This is the core design of the Niche Mortgages funnel: nine questions covering purpose, timeline, income shape and the rest, run through a scoring engine shared by the site and the server, so every lead arrives ALREADY scored and triaged: and the borrower gets something for their answers (a verdict on which tier of lender fits their file) before anyone calls. Compare that with the contact-form world the rebuild replaced, where every enquiry was a name and a phone number and the brokers spent their days qualifying instead of placing. One shared scoring engine matters more than clever weights: two scoring systems (one for the website, one in someone's head) drift into two truths.

The three lanes (and what each one gets)

• Hot lane (ready + reachable): the speed-to-lead promise lives here: same-business-hours callback, the file's funnel answers already on the broker's screen, the first call starting at question ten instead of question one.
• Warm lane (real file, longer clock): a scheduled, unhurried first conversation plus enrolment in the nurture machinery: renewal-date capture, the monthly note, calculator links. These borrowers transact on their calendar; the system's job is being present when it arrives.
• Nurture lane (someday or unclear): no call slot spent: the content library does the work. The Niche Mortgages engine (0 to 30+ articles a month per the case study, library up 70% from 46 to 78+ posts) exists partly for exactly this lane: education that keeps the brokerage the familiar answer until a someday becomes a timeline.

Closing the loop: scores must learn

• Calibrate monthly against outcomes: if hot-lane leads don't fund at a visibly higher rate than warm-lane ones, the weights are wrong: adjust as a team, briefly, using funded files as the referee.
• Feed scores to the campaigns: the NM Google and Meta campaigns are explicitly judged on applications, not raw lead count: that judgment is only possible because the funnel scores what each channel sends. A channel delivering high volumes of nurture-lane leads is a different purchase than one delivering hot-lane files at twice the CPL.
• Watch the lane mix as a health metric: a rising nurture share means targeting is drifting broad; a hot lane that never fills means the capture is asking the wrong questions or the traffic is wrong. The mix is a dashboard, not just a sorting rule.
• Keep the human override: a broker who smells a mis-scored lead reroutes it and notes why: those notes are next month's calibration input.

Start this week, upgrade forever

Day one is a shared sheet: source, purpose, timeline, 1-10 gut score, lane, outcome: filled at capture or first contact. The upgrade path runs through capture that scores itself (the funnel), a CRM that routes lanes automatically, and campaigns judged on what their leads became. The destination is the architecture on the case study page: every lead scored before the first hello, every broker hour aimed at the files that fund, every channel accountable for the quality it sends. Nova builds that system for brokerages end to end; triage is where it starts.