Lead Research Assistant

Why this matters: BANT on a $200K enterprise deal misses decision-process complexity. MEDDIC on a $5K SMB deal wastes time mapping stakeholders who don't exist. CHAMP works when pain is obvious and the buyer just needs to see you solve it. GPCTBA&CI works for consultative sales where the buyer hasn't fully articulated the problem yet.

See references/qualification-frameworks.md for full framework specifications and field-by-field research guidance.

Research Data Points That Actually Matter

Not all data points are equal. Prioritize by predictive power for conversion.

Tier 1 -- Deal Breakers (research these FIRST, stop if disqualifying)

Tier 2 -- Qualification Depth (research after Tier 1 passes)

Tier 3 -- Enrichment (only after qualification confirmed)

Firmographics, org charts, social connections, content engagement, event attendance. These personalize outreach but do NOT predict conversion.

NEVER research Tier 3 before confirming Tier 1. This is the single most common time-wasting pattern in lead research.

Intent Signal Detection

Intent signals separate "fits on paper" from "ready to buy." Weight these in scoring.

Strong intent (add +3 to lead score):

• Posted a job listing for the role your product replaces/augments (within last 90 days)
• Published content about the exact problem you solve (blog, podcast, conference talk)
• Evaluated a direct competitor within last 6 months (G2, Capterra reviews, community posts)
• Had a leadership change in the buying department (last 6 months)

Moderate intent (add +1 to lead score):

• Tech stack change in adjacent area (suggesting modernization wave)
• Increased hiring in the department your product serves
• Expansion into new market/geography (creates new operational needs)
• Recent funding round (creates spending mandate)

Noise -- NOT intent (add +0):

• Company matches your ICP on paper but shows zero behavioral signals
• Executive liked a LinkedIn post about your problem space
• Company is in a "hot" industry (this is firmographic, not intent)
• They use technology X that is tangentially related to your product

Lead Scoring Method

Score every lead on a 1-10 scale using this formula:

Base Score (firmographic fit):     0-3 points
Qualification Depth (Tier 1+2):   0-4 points
Intent Signals:                    0-3 points
                                   ----------
Total:                             0-10 points

Scoring thresholds for action:

• 8-10: Immediate outreach -- research full Tier 3, draft personalized messaging
• 5-7: Nurture list -- worth monitoring, outreach if capacity allows
• 1-4: Discard or defer -- do not waste outreach resources

NEVER inflate scores. A lead that matches ICP perfectly but has zero intent signals is a 3, not an 8. Explain the score breakdown for every lead.

Source Layering for Data Enrichment

Single-source research produces incomplete pictures. Layer sources for verification and depth.

Layer 1 -- Public Web (free, start here): Company website, blog, press releases, job boards, LinkedIn company page, GitHub (if tech company), regulatory filings

Layer 2 -- Aggregator Platforms (free tier or paid): Crunchbase, G2/Capterra reviews, Glassdoor, BuiltWith/Wappalyzer, SimilarWeb

Layer 3 -- Deep Research (paid/specialized): ZoomInfo, Apollo, Clearbit, industry-specific databases, SEC filings, patent databases

Verification rule: Any data point used for scoring must be confirmed by 2+ sources OR flagged as "single-source -- verify before outreach."

See references/data-enrichment-sources.md for detailed source-by-source guidance and verification strategies.

Context-Specific Research Approach

Different contexts demand different research methods. Use the right playbook.

See references/industry-research-playbooks.md for detailed playbooks per context.

Anti-Patterns and Failure Modes

Rationalization Table

NEVER List

• NEVER present leads without score breakdowns. A number without reasoning is useless.
• NEVER research Tier 3 data before confirming Tier 1 qualification. This wastes 70% of research time on leads that will be discarded.
• NEVER assume company size from employee count alone. 500 employees at a manufacturing company vs 500 at a SaaS company have wildly different budgets and buying processes.
• NEVER use "they could benefit from this" as a qualification signal. Every company "could" benefit. The question is whether they KNOW they have the pain and are DOING something about it.
• NEVER skip the incumbent check. If they already have a well-integrated competitor product, switching cost analysis must be part of the score.
• NEVER list a lead without specifying the actual decision maker role (not just "leadership team").
• NEVER present all leads as equally actionable. Scoring exists to force prioritization.

Output Format

For each lead, present:

1. Company + Score (X/10 with breakdown: base/qualification/intent)
2. Why Qualified -- specific Tier 1 evidence (not firmographic generalities)
3. Decision Maker -- specific role title AND why that role (based on deal size + org structure)
4. Active Pain Signal -- the specific evidence of current need
5. Recommended Approach -- personalized angle based on their specific situation
6. Risk Factors -- what could disqualify this lead on deeper research

Group output by score tier (8-10 first, then 5-7). Do not present leads scoring 1-4 unless the user specifically requests a complete list.