Realtor Lead Scraping and Verification Pipeline for Cold Email

Realtor Lead Scraping Pipeline

Pipeline Stages

1. Scraping Layer

• Use Zillow scraper (verified working)
• Target specific geos/service areas
• Cache results to avoid duplicate API calls across cities

2. ICP Verification

Filter for Ideal Customer Profile:

• Correct service area (instant cut if outside)
• Actual realtor (not assistant/team member)
• 99% confidence threshold

3. Deduplication Layers

Remove duplicates against:

• Existing leads in Superbase (previously scraped)
• Existing leads in Instantly campaigns (already being emailed)
• Existing client customers (from client's CRM export)

4. Data Normalization

• Standardize company names
• Clean phone numbers
• Normalize email formats
• Consistent field mapping

5. Email Verification

• Validate emails before adding to campaigns
• Remove bounced/invalid addresses
• Flag catch-all domains

Superbase Schema Requirements

Each client table needs:

• existing_customers column for dedup
• Standard lead fields (name, email, phone, company)
• source field (which scraper)
• scraped_date timestamp

Quality Control Process

1. Run small test batch (100 leads)
2. Run through full pipeline
3. Benchmark quality at end
4. If quality passes, run full service area scrape
5. If issues, debug before scaling

Proxy Management

• Use residential proxies for scraping
• Monitor credit usage
• Small batch test before full scrape to preserve credits

Campaign Integration

After pipeline completion:

1. Export verified leads
2. Import to Instantly AI
3. Assign to appropriate client campaign
4. Monitor bounce rates (pause if >5%)

Handling Multiple Clients

• Run pipelines for both clients together
• Each client has separate Superbase table
• Each client has separate Instantly campaigns
• Dedup runs per-client