Lead Score Model Setup

This drill designs a lead scoring model from scratch: define fit criteria (firmographics), intent signals (behaviors), assign point values, and manually score a batch of leads to validate the model predicts engagement.

Input

• ICP document (output from icp-definition drill)
• 20+ leads in Attio CRM with basic company and contact data
• PostHog tracking configured on your website (for intent signal capture)

Steps

1. Define fit criteria and point values

Using your ICP document, select 3-5 firmographic attributes that distinguish good-fit leads from poor-fit ones. Assign point values that sum to a maximum of 50:

| Fit Criterion | Condition | Points | |---------------|-----------|--------| | Company size | 20-500 employees (ICP range) | +15 | | Industry | Target industry match | +10 | | Buyer role | Decision-maker title (VP+, Head of, Director) | +15 | | Tech stack | Uses complementary/competing tool | +5 | | Geography | Target market | +5 |

Store these criteria as a scoring rubric in Attio notes on your lead scoring campaign record.

2. Define intent signals and point values

Select 3-5 behavioral signals that indicate purchase readiness. Assign point values that sum to a maximum of 50:

| Intent Signal | Condition | Points | |---------------|-----------|--------| | Demo request | Submitted demo/contact form | +20 | | Pricing page view | Visited pricing page (PostHog event) | +15 | | Email reply | Replied to outreach email | +10 | | Content download | Downloaded case study or whitepaper | +5 | | Repeat website visit | 3+ sessions in 14 days | +5 |

Note: At Smoke level, some intent signals may need manual observation (checking PostHog for website events, checking email tool for replies). Automation comes at Baseline.

3. Build the scoring formula in Clay

Using the clay-scoring fundamental, create a Clay table with your 20+ leads and add formula columns:

• Fit Score (0-50): Sum of fit criterion points based on enriched data
• Intent Score (0-50): Sum of intent signal points (manually populated at Smoke; automated later)
• Composite Score (0-100): fit_score + intent_score
• Tier: IF(composite >= 80, "Hot", IF(composite >= 50, "Warm", "Cold"))

4. Score the initial batch

For each of the 20+ leads:

1. Check enriched firmographic data against fit criteria. Assign fit points.
2. Check PostHog for website activity (use posthog-custom-events to query page views). Assign intent points.
3. Check email/outreach tool for reply history. Assign intent points.
4. Compute composite score and tier.

5. Push scores to Attio

Using the attio-lead-scoring fundamental, write fit_score, intent_score, lead_score, and lead_tier to each person record in Attio. Set last_scored to today's date.

6. Log scoring events to PostHog

Using posthog-custom-events, fire a lead_scored event for each lead with properties:

{
  "person_id": "...",
  "company_name": "...",
  "fit_score": 35,
  "intent_score": 25,
  "lead_score": 60,
  "lead_tier": "Warm",
  "scoring_method": "manual",
  "scoring_model_version": "v1"
}

7. Validate score distribution

Check that the score distribution is reasonable:

• At least 15-25% of leads should be Hot (if <15%, thresholds are too strict; if >40%, too lenient)
• Cold leads should represent the bottom 30-50%
• If all leads cluster in one tier, adjust point values to create separation

8. Document the model

Create an Attio note on the lead scoring campaign record with:

• Fit criteria and point values
• Intent signals and point values
• Tier thresholds
• Initial score distribution (X Hot, Y Warm, Z Cold)
• Date created and version number (v1)

Output

• A documented lead scoring model with specific criteria and point values
• 20+ leads scored and tiered in Attio CRM
• lead_scored events in PostHog for analysis
• Score distribution validated

Triggers

This is a one-time setup drill. Re-run when model criteria need updating (after analyzing which criteria correlate with closed deals).