Methodology

How Richie Reach Works: The Full Methodology

By Sayudh Mukherjee

This page covers how signals are captured, filtered, classified, scored, enriched, and delivered. It is the technical detail behind the nine-stage pipeline.

Stage 1 - Signal Capture

A cron-based monitor polls public LinkedIn engagement around client-defined keyword themes every 4 hours. Each captured event includes the LinkedIn profile URL, engagement type (comment, reaction, share), post context, and timestamp.

Deduplication runs on a LinkedIn URN hash. If the same person appears across multiple posts in the same week, the records are merged to one with the engagement count incremented.

The monitor runs on our infrastructure. No client account access is required.

Stage 2 - ICP Gate

Records enter the ICP gate before any enrichment spend.

Hard gates (role, geography, company size) drop non-matching records immediately. Records that fail hard gates are not enriched. Soft gates (industry, funding stage, tech stack) reduce the composite score but do not automatically drop the record.

The gate runs on publicly available company and person information. Enrichment only starts after a record clears the gate. This is the primary cost control mechanism.

Stage 3 - AI Classification

An AI model reads the engagement text and assigns a behavioral intent class.

Class	What it means
evaluating	Comparing options or asking for recommendations
active_buyer	Mentions timeline, budget, or switching decision
category_engaged	Thoughtful engagement, no evaluation language
surface_engagement	Generic positive reaction, minimal signal value
disqualified	Job posts, self-promotion, off-topic content

The class is used in composite scoring and determines which outreach copy framework is applied.

Stage 4 - Composite Scoring

Records are scored from 0.00 to 1.00. Inputs: engagement class weight, ICP fit score, signal recency (older signals score lower), engagement depth (comment vs reaction, length of comment).

Records scoring 0.74 and above: HIGH. Delivered first. Records scoring 0.52 to 0.73: MED. Delivered. Records scoring below 0.52: not enriched, not delivered.

Stage 5 - Enrichment Waterfall

Only records that clear the ICP gate and score at or above 0.52 enter the enrichment waterfall.

Three tools run in sequence: Prospeo (primary, approximately 83% match rate), Apollo (fallback), Hunter (domain pattern, last resort). If all three miss, the record is delivered with LinkedIn URL only and email set to null.

Approximate cost per enrichment run on a filtered set of 87 records: $0.87. On an unfiltered set of 500 records: $5.00. The gate drives the cost difference.

Stage 6 - Delivery

Enriched records are delivered weekly via CSV, Clay table, or webhook to Smartlead, Lemlist, Instantly, HubSpot, or a custom API endpoint. Every record includes the full signal metadata so context travels with the contact into your outreach tools.

Stage 7 - CRM Sync and Feedback Loop

If HubSpot is in the stack, all active records sync with signal metadata as contact properties. When a deal closes, the signal theme, engagement class, and score at the time of capture are available in the CRM.

This attribution data is used to refine which keyword themes and score thresholds produce real pipeline in your specific market, not just replies.

See the full interactive pipeline on the How It Works page.