Lead Qualification

Qualification protects selling time. Score honestly, disqualify fast, document why. Use the deterministic helper for ICP totals and CSV validation.

The live book is Studio #/leads. leads action=rescore writes BANT-F on the desk. The desk loop and leads action=watch re-score locally. Heartbeat never hunts and never sends a sequence.

When to use

• Ranking a prospect list before outreach
• Routing A/B/C after enrichment or discovery

When not to use

• Pure research sheets without scoring (account-research)
• Inflating scores to hit activity quotas

Scoring model (BANT-F + ICP)

Dimension	Questions	Signals
Budget	can they pay?	size, funding, current spend
Authority	decider or tourist?	role, buying process
Need	real pain?	trigger, cost of inaction
Timing	why now?	deadline, renewal, project
Fit / ICP	can we serve them?	sector, size, geo, stack

Score each dimension 0-5 with evidence. Weighted default: Fit 30%, Need 25%, Timing 20%, Authority 15%, Budget 10%.

Tiers from total 0-100:

Tier	Range	Action
A	≥70	contact now
B	40-69	nurture
C	<40	discard / revisit condition

Disqualification triggers

• No identifiable pain we solve
• Budget an order of magnitude off
• Fit failure (out of ICP)
• Deciders unreachable after agreed attempts

Log reason + revisit condition ("re-check after FY").

Helper script

# Validate CSV columns / URLs / confidence
python navin/skills/lead-qualification/scripts/score_leads.py sales/prospects.csv --validate-only

# Score rows that already have bant columns (fit,need,timing,authority,budget)
python navin/skills/lead-qualification/scripts/score_leads.py sales/prospects.csv -o sales/prospects-scored.csv

Expected optional columns for scoring: fit,need,timing,authority,budget (0-5 each) or a single icp_score.

Workflow

1. Input leads + discovery/enrichment notes.
2. Fill BANT-F; unknowns become next-call questions (do not invent).
3. Run the script; route A/B/C.
4. Update after material changes; pipeline-analyst consumes tiers.

Rules

• Every score cites evidence or stays unknown.
• Optimism is not a data point.
• Never mark email verified without enrichment proof / public source.

Anti-patterns

• Scoring everyone A
• Dropping disqualified rows without reason codes