How contact scoring and tier labels are calculated

Contact scoring combines individual contact signals, such as job titles, with account-level signals to calculate a final score. This combined score determines whether a contact is labeled as Warm, Hot, or Burning based on your organization's percentile thresholds.

Context

When viewing your contacts, you might notice that contacts with the exact same job title are assigned to different tier labels (such as Warm and Hot). This occurs because contact scoring is not based solely on the contact's title.

How Contact Scoring Works

Your contact score is a combined value calculated from both contact-specific signals and account-level signals:

• Combined Signals: A contact's score is the sum of their individual title-signal points and all associated account-level signals (such as practice subtype, provider count, or other account-level focus areas).
• Percentile-Based Tiers: The Warm, Hot, and Burning labels on the contacts tab are computed from this combined score against your organization's percentile thresholds. Because two contacts with the same job title may belong to different accounts with different account-level scores, their final combined scores can differ and place them into different tiers.

How Unlisted Titles are Handled

If a contact's job title is not explicitly listed in your scoring rubric, Monaco handles it using semantic matching:

• Semantic Matching: Monaco uses a model-based semantic match rather than an exact-string match. It evaluates your rubric's bucket descriptions and examples to assign the contact to the closest matching level.
• Fallback Score: If a job title is completely unrelated to any of the categories in your rubric, it falls back to a score of zero.
• Updating Rubrics: If you want to ensure a specific title is always categorized into a particular bucket, add that title explicitly to the desired bucket in your rubric. Editing the rubric invalidates cached title matches, and contacts will receive the updated scores on the next scoring run.