← Archive

2024

Gov-Program Lead Crawler & Cold Email Engine

Automated B2B Lead Collection & Personalized Outbound Pipeline

Role	Solo Builder / Product Operations
Context	Internal GTM Automation Tool
Stage	Production Internal Workflow
Time	2024.11 – 2024.12

Summary	Automated lead collection and personalized cold-email pipeline targeting government startup program cohorts.
Scope	Reduced manual B2B lead research and outreach preparation time by 95%.
Outcome	95% reduction in manual lead research & outreach time
Tags	Internal Tool, Operations Automation, B2B Outbound, Python

01

Problem

Capability thesis	Replaced repetitive manual sales prospecting with a deterministic Python scraping, enrichment, and outbound dispatch pipeline.
TL;DR	Built an internal automation tool that crawls public government-grant startup cohorts, enriches company contact emails, and dispatches personalized B2B cold emails—reducing manual prospecting time by 95%.
What	An end-to-end Python pipeline combining web crawling, contact deduplication, domain/email verification, and batch cold-email generation.
Why now	Early-stage B2B validation requires reaching hundreds of funded startups quickly, but public cohort lists only publish company names without structured contact details.
Problem	Manually searching each selected startup on Google/Naver, visiting their homepage footer, copying their email into a spreadsheet, and sending individual emails took 5–10 minutes per lead and limited daily outbound volume.

02

Role & ownership

Role	Product Manager & Automation Developer (Solo)
Team	Solo execution
Ownership	Designed the data schema, wrote the crawler/enrichment scripts, implemented bounce/duplicate filtering, and operated the outbound email campaigns.
Constraints	Unstandardized HTML/PDF cohort announcements, rate-limiting on search engines, and email deliverability/spam-filter constraints.

03

Validation loops

01

Supported

Loop 1 — Automated Contact Enrichment & Dispatch Pipeline

Automating homepage discovery and footer/contact email extraction from government grant cohort lists will cut manual prospecting time by over 90%.

Design & Build

Why	Remove repetitive browser tab switching and copy-pasting.
Actions	Built a Python crawler that resolves company domains, extracts valid contact emails, deduplicates records, and merges cohort context into outbound templates.
Outputs	Reusable CLI/script pipeline for B2B lead generation.

Run

Why	Power live B2B validation outreach.
Actions	Executed automated enrichment and cold-email batches across startup cohorts.
Outputs	Cut manual lead-prep time by 95% and fueled 2,000+ B2B outbound contacts.

04

Design decisions

Problem zoom	Broke the manual workflow into four automatable stages: (1) Cohort entity extraction, (2) Official homepage resolution, (3) Contact email regex/DOM scraping, and (4) Personalized message merge.
Options	(1) Hire part-time data-entry freelancers, (2) Buy expensive static B2B databases with outdated startup contacts, (3) Build a custom crawler targeting freshly funded startup cohorts.
Trade-off	Prioritized high-precision extraction with fallback flags for ambiguous domains over blind scraping, preventing embarrassing mis-sends.
Final rules	Deduplicate by domain and email address, log source URLs for every extracted contact, and throttle outbound sends with randomized intervals.
Tools	Python, Playwright / BeautifulSoup, Regex, CSV/Google Sheets, SMTP automation
Frameworks	ETL Pipeline Design, Outbound Sales Ops Automation

05

Outcomes

Metrics	Reduced per-lead research and email preparation time by ~95%, enabling 2,000+ targeted B2B outreach touchpoints without additional headcount.
Qualitative outcome	Turned B2B lead generation from a tedious manual chore into a repeatable on-demand script reusable across multiple product validation loops.

06

Learnings

Learnings	Automating the top of the B2B funnel frees founder bandwidth to focus on high-leverage discovery calls and product iteration.
If again	Add an LLM classification pass on each crawled landing page to tag the startup's industry vertical and recent hiring roles automatically.

07

Evidence

Before	Manual search-and-copy prospecting taking hours per 50 companies.
After	Automated cohort-to-inbox pipeline cutting manual lead ops time by 95%.