EdTech SaaS: Rebuilding Pipeline Architecture from Scratch

The Client

A Series B EdTech SaaS company with 80 employees, a 12-person sales team, and a HubSpot CRM that had grown organically for three years without architectural oversight.

The Problem

The VP Sales couldn't trust the pipeline numbers. Forecast reviews were arguments, not analysis. The specific failures:

• 450+ custom properties, fewer than 40 in active use
• No lifecycle stage definitions — MQL meant "marketing touched them" to some reps and "they requested a demo" to others
• Lead routing by round-robin only — no priority scoring, no territory logic, no fallback chains
• No attribution — marketing couldn't prove which campaigns generated pipeline
• Three pipeline views that contradicted each other because they used different deal stage definitions

The result: forecast accuracy was below 50%. The board was losing confidence. Sales and marketing blamed each other for pipeline quality.

The Engagement

Month 1: Pipeline Audit

We ran a full pipeline audit — a diagnostic that maps the entire revenue system from lead capture to closed-won.

Key findings:

• 62% of deals in "Qualified" stage had no qualification criteria met
• Average speed-to-lead was "4 hours" but the median was 22 hours (skewed by a few fast responses)
• Marketing attribution was impossible because UTM parameters weren't mapped to CRM properties

Month 2: Object Model Redesign

We redesigned the CRM data model:

• Reduced properties from 450 to 85 (all documented, all with owners)
• Implemented predicate-based lifecycle stages: MQL = fit score ≥ 40 AND engagement score ≥ 30
• Built deal stage gates: can't move to "Demo Scheduled" without a meeting on the calendar

Month 3: Routing & Automation

We rebuilt lead routing:

• Priority scoring by fit + engagement + intent signals
• Territory-based primary assignment with cross-territory fallback
• 2-hour SLA with auto-escalation to backup rep
• Manager alert at 4-hour breach

Month 4: Attribution & Reporting

We connected the pipeline to campaign attribution:

• UTM → CRM field mapping
• Multi-touch attribution model (first-touch + last-touch + linear)
• Pipeline contribution dashboard by campaign, channel, and offer

The Results

Metric	Before	After
Forecast accuracy	48%	94%
Speed to lead (median)	22 hours	1.4 hours
MQL → SQL conversion	12%	31%
Pipeline visibility	3 conflicting views	1 source of truth
Properties in use	40 of 450	85 of 85

Key Takeaway

The CRM wasn't broken — it had never been built. Three years of ad-hoc configuration had created complexity without structure. The fix wasn't a new tool — it was architecture.

---

Related Products

• Pipeline Audit — The diagnostic that started this engagement
• CRM Architecture — Object model and schema redesign
• Revenue Systems — End-to-end revenue infrastructure

→ Book a Consultation — If this sounds familiar, let's talk.