Back to Help Center

Correlation engine

What the correlation engine does

Individual signals are useful. Correlated signals are predictive. The Revnox correlation engine continuously scans for companies exhibiting multiple signals simultaneously, then scores and surfaces them as high-priority accounts.

How correlation detection works

Step 1: Signal collection

Every 6 hours, Revnox collectors pull fresh data across all 14 signal categories for your watched companies.

Step 2: Pattern matching

The engine compares each company's active signals against a library of known high-intent patterns:

  • "Growth hire" pattern: Funding + SDR hire + VP Sales hire within 60 days
  • "Vendor switch" pattern: Competitor complaint + tech tool removed + tech tool added
  • "Executive reset" pattern: New CEO or CRO + departures in leadership team
  • "Product-market push" pattern: Product launch + marketing hire + content spike

Step 3: Score multiplication

When a pattern matches, the account's ICP score is multiplied:

  • 2 correlated signals: 1.5× multiplier
  • 3 correlated signals: 2.5× multiplier
  • 4+ correlated signals: 3× multiplier

This pushes high-correlation accounts to the top of your signal dashboard even if individual signal strengths are moderate.

Viewing correlations

In the Signals view, correlated accounts show a purple "Correlated" badge with a count of matching signals. Click through to see the full correlation breakdown and the specific pattern detected.

Custom patterns

On Scale tier, you can define custom correlation patterns in Settings > Signal Patterns. Useful if your ICP has unique triggers that don't map to the default library.

AI Signal Intelligence for GTM Teams | Revnox