Lead Scoring

Lead scoring ranks leads by conversion likelihood so teams can prioritize outreach and qualification effort.

Goals

score_total is calculated on a 0-100 scale from weighted signals:

Formula:


score_total = clamp(0, 100, fit_score + intent_score + engagement_score - negative_score)

Category	Weight	Example factors
Fit	40	Company size match, industry match, geography, job title
Intent	35	Pricing page visit, demo request, high-value form answers
Engagement	25	Email open/reply, call connected, repeat site visits
Negative	-30 max	Bounce, unsubscribe, invalid contact, spam pattern

Weights are tenant-configurable with sane defaults.

Band	Range	Meaning
`hot`	70-100	High-priority follow-up now
`warm`	40-69	Qualified for nurture or light-touch follow-up
`cold`	0-39	Low-priority; monitor or disqualify

Band thresholds are configurable per tenant.

Score recalculation runs on:

Stores explainability and audit trail for each recompute.

Column	Type	Notes
`tenant_id`	uuid FK → tenants
`lead_id`	uuid FK → leads
`model_version`	varchar(50)	e.g. `v1.0`
`factor_key`	varchar(100)	e.g. `job_title_match`
`factor_category`	enum	`fit` \| `intent` \| `engagement` \| `negative`
`points`	integer	Signed points contribution
`details`	jsonb	Raw evidence for the factor
`created_on`	timestamptz

Indexes:


CREATE INDEX idx_lead_score_factors_lead
  ON lead_score_factors(tenant_id, lead_id, created_on DESC);