How Lead Scoring Works

The Scoring Model

Trust Leads's lead score is a rule-based composite score computed from a weighted set of validation signals. It is deliberately not a machine-learning model — this means it is fully deterministic, auditable, and predictable. You can understand exactly why any given contact received its score by examining the enrichment_flags column.

The score starts at a baseline of 100 and deductions are applied for each failing signal. Positive bonuses are added for extra-strong signals (such as a corporate email domain with a perfect MX record and a known format). The final score is clamped between 0 and 100.

Positive Score Signals

Several signals increase a contact's score when they are present and verified. A confirmed deliverable email (syntax valid + MX found + not disposable) contributes the largest positive weight. Name completeness (both first and last name present and non-numeric) adds to the score. A valid, normalised phone number adds to the score.

Domain-level signals can also boost the score: a corporate email domain (not a free provider like Gmail, Yahoo, or Outlook) suggests a real business contact and contributes a positive signal. DMARC and SPF records on the domain add a small bonus, indicating the domain is actively managed.

• Email syntax valid: +15 points
• Email MX record confirmed: +25 points
• Email not disposable: +10 points
• Corporate (non-free-provider) email domain: +10 points
• Phone number valid and normalised: +10 points
• First and last name present: +10 points
• Domain has DMARC record: +5 points
• Company name present: +5 points
• Job title present: +5 points

Negative Score Signals

Negative signals are applied as deductions from the accumulated positive total. The harshest deductions are reserved for signals that strongly indicate a contact will never be useful: disposable email domains and completely invalid email syntax both cause large deductions.

Moderate deductions are applied for signals that significantly reduce deliverability or confidence: no MX record on the email domain, a suspicious TLD, a phone number in an unrecognised format. Small deductions apply to minor data quality issues like a missing job title or company name.

• Disposable email domain: -50 points
• Email syntax invalid: -40 points
• No MX record on email domain: -25 points
• Suspicious TLD: -20 points
• Domain typo detected: -15 points
• Invalid phone format: -10 points
• Missing first or last name: -5 points each

The enrichment_flags Column

Every contact gets an enrichment_flags column in the output CSV. This column contains a pipe-separated list of all the flags triggered during processing. The value 'OK' means no negative signals were found. Flags are in SCREAMING_SNAKE_CASE for easy programmatic parsing.

You can use the flags column to build targeted remediation workflows. For example, filter on DOMAIN_TYPO and send those contacts to a data hygiene workflow to verify the correct email. Filter on NO_MX_RECORD and route those to a manual research queue to find the contact on LinkedIn.

# Common enrichment flag values
OK                   — No issues found
INVALID_EMAIL        — Email fails syntax validation
NO_MX_RECORD         — Domain has no MX DNS records
DISPOSABLE_EMAIL     — Email domain is on disposable blocklist
DOMAIN_TYPO          — Possible typo in email domain detected
SUSPICIOUS_TLD       — Email domain uses a high-risk TLD
INVALID_PHONE        — Phone number present but fails format check
MISSING_PHONE        — No phone number in the input record
MISSING_NAME         — First or last name (or both) not present