Teardown: A Lead Gen Funnel That Produces 1,200+ Contacts a Month

The funnel, in four parts

One client system we built generates 1,200+ qualified contacts per month on autopilot. The architecture isn't exotic — capture, qualify, nurture, deliver — but each part has one decision that separates funnels that produce from funnels that exist. This teardown is about those decisions.

Part 1 — Landing pages built for one intent each. Not a website — a page, per intent: one for valuation seekers, one for first-time buyers, one per farm neighborhood. The conversion killer in most funnels is a page trying to serve three visitors at once. One page, one promise, one form. Traffic source and page intent must agree, or you're paying to confuse people.

Part 2 — Qualification is staggered, not front-loaded. The instinct is to qualify at the form — ask timeline, budget, financing up front. Every added field costs conversion, and the leads you priced out of asking don't come back. The working pattern: capture on the minimum (contact plus the one intent question the page already implies), then let the sequence qualify — each automated touch asks one natural next question, and the profile assembles over days. Qualification is a conversation, not a toll booth.

Part 3 — Nurture matched to timeline. The sequence forks on what the lead's answers reveal: near-term intent accelerates to a human handoff; six-months-out goes to a patient drip that stays useful (market updates, neighborhood data) instead of needy. The expensive mistake is treating every lead as now-or-never — most of a funnel's eventual revenue sits in the long-timeline cohort that lazy funnels burn out or ignore.

Part 4 — Delivery into the pipeline, with provenance. Every contact lands in the CRM carrying source, page, campaign, and its accumulating answers — feeding the speed-to-lead machinery and, just as important, feeding measurement. Cost-per-qualified-lead by source is the number that makes scaling a budget decision, and it only exists if provenance survives the trip.

Where the first version broke

Attribution silently died and the funnel looked like a money pit. A tracking-parameter change upstream broke source attribution mid-month; leads kept flowing but arrived source-less. Dashboards showed one channel's cost-per-lead spiking toward infinity while contacts piled into an "unknown" bucket nobody watched. The near-miss: budget was almost pulled from the funnel's best channel. Since then, attribution has a heartbeat — unknown-source share is an alarmed metric, because a funnel you can't measure is a funnel you'll defund by accident.

A form asked one question too many. Adding a phone-number requirement to a page that had been email-only cut conversions sharply enough to see in a week. It wasn't the field — it was the trade: the page's offer wasn't worth a phone number yet, three touches later it was. Every field earns its place by paying for its conversion cost, and the ledger gets checked with data, not opinion.

The full teardown — the page-per-intent structure, the staggered qualification sequence design, the timeline-fork logic, and the attribution heartbeat setup — lives in the Builder library.