---
title: "Un-inverting a funnel and rebuilding attribution across a ~285k-contact database"
type: case-study
evidence_type: proof
category: "Attribution & Lead Lifecycle"
publisher: "LeanScale"
date_modified: 2026-08-08
word_count: 401
topics: ["demand-generation", "revenue-operations"]
canonical_url: https://knowledge.leanscale.team/customers/un-inverting-a-funnel-and-rebuilding-attribution/
source: "LeanScale Knowledge Hub — https://knowledge.leanscale.team"
license: "Free to quote and cite with attribution to LeanScale."
---

# Un-inverting a funnel and rebuilding attribution across a ~285k-contact database

**Evidence type:** proof (what happened)

A construction-industry software company running a mixed self-serve and sales-led motion had a funnel where Opportunity counts ran larger than SQL and ~82% of contacts sat in an unattributed bucket. LeanScale rebuilt the lifecycle model with forward-only transitions, built a 14-channel attribution framework, and stamped GTM motion across the entire contact base.

## The challenge

The funnel was inverted: Opportunity counts sat 3–7x larger than SQL because lifecycle stages never regressed when deals closed, so no conversion rate and no velocity number could be trusted. Roughly 82% of contacts were bucketed into an unattributed 'offline' / integration source, which left marketing unable to say what any channel had produced. There was no channel taxonomy, no UTM hygiene, no hidden-field capture on forms, and no way to separate inbound from outbound at all.

## The approach

GTM diagnostic first
Ran a full go-to-market diagnostic across the funnel before changing anything, to establish exactly where records were stranded, which transitions were firing backwards, and what share of the database was unattributable.

Lifecycle rebuild with forward-only logic
Rebuilt the lifecycle stage model end to end with forward-only transition logic and velocity stamping, then re-stated every stranded record so historic conversion and velocity became computable.

Attribution framework
Built a 14-channel taxonomy with normalized 'golden' attribution fields, an eight-step attribution workflow engine, UTM hygiene and hidden-field capture on forms, and webinar attribution that recovered ~3,300 previously invisible attendees.

Motion stamping across the whole base
Stamped and backfilled lead motion — inbound, outbound, other — across the entire contact base, and stood up separate inbound and outbound lead-status engines so the two motions stopped contaminating each other's metrics.

Channel ROI and forecasting
Stood up the Vasco GTM data platform to fuse attribution with channel spend for CAC and ROI reporting and segment-level forecasting, built a first-version growth model, and delivered the company's first formal quarterly business review with a top-10 metrics readout.

Identity resolution across product, CRM and billing
Stitched product, CRM and billing data into a canonical user-to-organization-to-revenue identity table, so usage-based upsell signals could be attached to the right revenue-bearing account.

## Outcomes

98% lifecycle-stage accuracy
After the rebuild and backfill, 98% of contacts and companies sit in the correct stage, fixing a funnel where Opportunity had run 3–7x larger than SQL.

~285k contacts attributed to a GTM motion
Lead motion was stamped and backfilled across 100% of the contact base, and channel plus record source now populate ~99.9% of contacts.

~3,300 previously invisible attendees attributed
Webinar attribution captured event attendees who had never been connected to a channel or campaign in the CRM.

Channel spend joined to outcomes
Attribution was fused with channel spend in a GTM data platform for CAC / ROI and segment-level forecasting, feeding the company's first formal quarterly business review.

## Quotes

> we would like you to project manage the improvements to our data infrastructure and we would like to move on it ASAP
>
> — Customer, CTO


## Canonical

https://knowledge.leanscale.team/customers/un-inverting-a-funnel-and-rebuilding-attribution/