From a seed list to a qualified outbound database

Problem

Every new outbound campaign needed a fresh database of companies and decision-makers. Building it by hand was slow, hard to repeat across clients, and easy to overspend on enrichment credits.

What I built

A deterministic Python pipeline configured per client and ICP. It ingests a seed list, enriches companies with traffic data (Apify), finds people with Apify and an Apollo fallback, and filters both against ICP rules. A probe mode estimates the universe and cost before any credits are spent, and every run is resumable and auditable (raw snapshots, a report, and a state file).

Result

Validated end-to-end on a 100-company run. A full run for one ICP was estimated at about $5 plus ~200 Apollo credits. Onboarding a new client is a config change, not a rebuild.