SERP Query Construction

How Enrich Engine builds search queries for LinkedIn discovery.

Understanding how search queries are constructed can help you optimize your CSV inputs for better match rates. This guide explains the query patterns used by Enrich Engine.

Query Patterns

Role + Company Search

The most common pattern for finding people by title at a company:

site:linkedin.com/in/ "VP of Sales" "Anthropic"

This searches for LinkedIn profile pages containing both the role and company name.

Name + Company Search

For finding specific people by name:

site:linkedin.com/in/ "John Smith" "Anthropic"

Adding company context significantly improves accuracy for common names.

Name Only Search

Without company context:

site:linkedin.com/in/ "John Smith"

Less accurate for common names. Works well for unique or distinctive names.

Company Domain Search

To find a company's official website from their name:

site:www. "Anthropic"

Used internally to resolve company names to domains when needed.

Result Parsing

LinkedIn URL Extraction

The system looks for URLs matching this pattern:

https://linkedin.com/in/[username]
https://www.linkedin.com/in/[username]

Name Parsing from Titles

LinkedIn search results have titles in this format:

John Smith - VP of Sales at Anthropic | LinkedIn

The parser extracts names by:

  1. Removing "| LinkedIn" or "- LinkedIn" suffix
  2. Taking text before the first " - " separator
  3. Handling comma-separated credentials (e.g., "John Smith, MBA")
  4. Filtering out common titles (Dr, Mr, Mrs, PhD, etc.)
  5. Taking first word as first name, last word as last name

Skipped Domains

When searching for company websites, these domains are automatically skipped:

linkedin.com
facebook.com
twitter.com
x.com
instagram.com
youtube.com
wikipedia.org
crunchbase.com
glassdoor.com
indeed.com
bloomberg.com
forbes.com
reuters.com
zoominfo.com
apollo.io
pitchbook.com
g2.com
yelp.com
bbb.org

This ensures we find official company websites, not profiles on aggregator sites.

Optimization Tips

Exact Match Quotes

All search terms are wrapped in quotes for exact matching. "VP of Sales" matches the exact phrase, not pages with "VP" and "Sales" separately.

Site Restriction

site:linkedin.com/in/ restricts results to LinkedIn profile pages only, excluding company pages and other LinkedIn content.

Result Ordering

Google ranks results by relevance. The first LinkedIn profile URL in results is typically the best match. For multi-person searches, subsequent results are used for additional matches.

Deduplication

When scraping multiple rows with the same company/role, previously found URLs are excluded to ensure unique results. This is handled via aroleInstanceIndex that tracks which result to use.

Common Issues

No Results Found

  • Person may not have a LinkedIn profile
  • Profile has privacy settings preventing indexing
  • Company name doesn't match LinkedIn company page
  • Job title is too generic or doesn't exist at the company

Wrong Person Found

  • Common name without company context
  • Person with similar name at same company
  • Company name is ambiguous (subsidiaries, similar names)

Improving Match Rate

  • Use official company names as shown on LinkedIn
  • Add company context for name searches
  • Use specific job titles, not generic terms
  • For very common names, consider adding middle initial