Resolution

When you license the Signal Database, every signal record comes pre-resolved with multiple identifiers. This page explains how resolution works and how to efficiently match signals to your existing data.

How Resolution Works

Unlike the real-time API (which resolves entities on-demand from minimal input), the Signal Database delivers pre-resolved records with all available identifiers populated.

What We Resolve

For each signal, we resolve and include:

Contact-Level Signals:

• Professional email address
• LinkedIn profile URL
• Current company domain
• Current company LinkedIn URL

Company-Level Signals:

• Primary company domain
• Company LinkedIn URL
• Normalized company name

Resolution Sources

We resolve entities by cross-referencing:

• LinkedIn profile data
• Email domain verification
• Company website analysis
• Employment history correlation
• Public business registrations

Matching Signals to Your Data

Recommended Join Keys

Use these fields in order of reliability:

Best Practices

For Contact Matching:

-- Example: Join signals to your contacts table
SELECT 
  c.id as contact_id,
  s.*
FROM signals s
LEFT JOIN contacts c 
  ON LOWER(s.entity_identifiers.contact_email) = LOWER(c.email)
  OR s.entity_identifiers.contact_linkedin_url = c.linkedin_url
WHERE s.entity_type = 'contact'

For Company Matching:

-- Example: Join signals to your accounts table
SELECT 
  a.id as account_id,
  s.*
FROM signals s
LEFT JOIN accounts a 
  ON LOWER(s.entity_identifiers.company_domain) = LOWER(a.domain)
  OR s.entity_identifiers.company_linkedin_url = a.linkedin_url
WHERE s.entity_type = 'company'

Handling Unmatched Signals

Not every signal will match a record in your database. Unmatched signals are still valuable:

1. New prospects: Signals on contacts/companies not yet in your CRM represent potential new leads
2. Data enrichment: Use unmatched signals to expand your TAM with pre-qualified targets
3. ICP refinement: Analyze which signal types appear for contacts outside your database

Coverage Expectations

Resolution coverage varies by identifier type:

Deduplication

Signals are deduplicated before delivery:

• Same signal, same entity: Only the most recent detection is included
• Refresh window: Signals are refreshed based on their category (see Delivery for cadence)
• Entity merging: If we detect that two records represent the same entity, we merge and deliver under the canonical identifiers

Data Quality

Validation Rules

All signals pass validation before delivery:

• Email addresses are format-validated and domain-verified
• LinkedIn URLs are normalized to canonical format
• Company domains exclude generic providers (gmail.com, etc.)
• Timestamps are standardized to ISO 8601 UTC

Confidence Indicators

For some signal types, we include confidence metadata in the variables object:

{
  "variables": {
    "resolution_confidence": "high",
    "last_verified_at": "2024-12-01T00:00:00Z",
    // ... other fields
  }
}

Comparison: API vs. Signal Database Resolution

Questions?

If you have specific resolution requirements or need custom identifier mapping, contact us at [email protected].