PostHog Trial Conversion Intelligence

Step 1 — The Fetcher

Code-only

Retrieves trial user event data from PostHog API — page views, feature usage events, session counts, unique days active, team invitations, and integration activations. Filters to users within the configurable trial window and assembles per-user behavioral profiles with event counts and timestamps.

Step 2 — The Assembler

Code-only

What does The Assembler actually decide? Computes 5 PQL (Product Qualified Lead) dimensions for each trial user: activation_depth (30%) — completion of key activation milestones, usage_velocity (25%) — session frequency and recency trend, feature_breadth (20%) — count of distinct features used vs power feature list, team_signal (15%) — team invitations, shared workspaces, collaborative events, engagement_recency (10%) — days since last activity and session trend direction.

Step 3 — The Scorer

Tier 2 Classification

This step exists because raw data alone is not enough. Computes weighted PQL score (0-10) and classifies each trial user: HOT (≥8) — ready for sales touch, WARM (5-7) — nurture with product-led content, COLD (<5) — low engagement, deprioritize or product intervention needed. Generates per-user conversion strategy with specific feature adoption gaps and recommended next actions.

Step 4 — The Writer

Tier 3 Creative

The output here feeds everything downstream. Generates personalized conversion outreach email drafts per PQL tier. HOT users get direct sales outreach with pricing and plan options. WARM users get feature discovery content based on unused power features. COLD users get re-engagement content highlighting the features they used most. Each draft references specific usage data.

Step 5 — The Formatter

Code-only

Without this step, upstream analysis sits idle. Creates Gmail draft emails for HOT and WARM users, creates Pipedrive deals for HOT users (if not already tracked), and posts a daily Slack summary with PQL distribution, conversion funnel metrics, top 5 HOT users, and feature adoption gaps across the trial cohort.n8n nests webhook data under a body key. Two lines of defensive code — const data = raw.body || raw — prevent the most common deployment failure.