Lead Scoring Framework Builder
Build an ICP-based lead scoring model that actually predicts conversion. Stop arguing about MQL definitions with a data-driven framework.
Most lead scoring models don't survive contact with the sales team. Teams ship 30-signal models that nobody trusts, and reps end up ignoring the "MQL" flag entirely. What actually works in production is a simpler two-axis grid — ICP fit on one axis, engagement on the other — with thresholds calibrated against real closed-won cohorts rather than picked on a whiteboard.
See why we fix your CRM before we spend and enterprise GTM fails when you measure.
The Two-Axis Scoring Model
Effective lead scoring has two independent components:
| Axis | What It Measures | Max Points |
|---|---|---|
| ICP Fit Score | How well does this lead match your ideal customer profile? | 50 points |
| Engagement Score | How actively is this lead engaging with your content/product? | 50 points |
Why two axes? A perfect-fit company with no engagement isn't ready for sales. A highly engaged lead from a bad-fit company will waste sales time. You need both to be high.
Where Intent Signals Actually Come From
"Engagement" as a scoring axis is too narrow if you only measure it from forms and email clicks. By the time a prospect fills out a demo form, they've already done 60-70% of their research elsewhere. Broader scoring pulls from five signal sources — each with a different timing window relative to the sales conversation.
| Signal Source | Examples | What it tells you | Timing before sales-ready |
|---|---|---|---|
| Community & Social Listening | Reddit threads, Slack communities, G2/Capterra reviews, LinkedIn discussions | Unfiltered buyer pain in their own language | 4-6 months |
| Research & Intent Data | ZoomInfo Insights, Bombora, 6sense, Breeze intent scores | Company-level research activity | 1-3 months |
| Behavioral (First-Party) | Pages viewed, time on site, content downloads, email engagement | Individual-level intent | Days to weeks |
| Social Engagement | LinkedIn profile views, content engagement, webinar attendance | Early awareness and interest from named personas | 2-3 months |
| Anonymous Signals | ICP-company traffic (no contact), relevant job postings, tech stack changes | Company-level buying triggers before anyone identifies themselves | 3-6 months |
If your engagement score only includes form fills and email clicks, you're scoring the tail end of the journey. Most pipeline-predictive intent shows up weeks or months earlier — in community conversations, anonymous site visits from target accounts, and job postings. Score those too, even if the lead isn't "known" yet.
See our intent signals playbook for how to wire anonymous signals into HubSpot without making the scoring model unreadable.
ICP Fit Scoring (50 points max)
Score these firmographic and demographic attributes:
| Attribute | Ideal Match | Points |
|---|---|---|
| Company Size | Your sweet spot (e.g., 200-2000 employees) | +15 |
| Adjacent (e.g., 50-199 or 2001-5000) | +8 | |
| Outside range | 0 | |
| Industry | Primary target industry | +15 |
| Secondary industry | +8 | |
| Non-target | 0 | |
| Job Title/Seniority | Decision maker (VP+) | +10 |
| Influencer (Manager/Director) | +6 | |
| Individual contributor | +2 | |
| Geography | Primary market | +5 |
| Secondary market | +2 | |
| Tech Stack | Uses complementary tech | +5 |
Engagement Scoring (50 points max)
Score behavioral signals by intent strength:
| Action | Intent Level | Points |
|---|---|---|
| Demo/Trial Request | High | +25 |
| Pricing Page Visit | High | +15 |
| Case Study Download | High | +10 |
| Product Page (2+ visits) | Medium | +8 |
| Webinar Attendance | Medium | +8 |
| Whitepaper/Guide Download | Medium | +5 |
| Blog Visit (3+ articles) | Low | +3 |
| Email Click | Low | +2 |
| Email Open | Low | +1 |
Score Decay: The Missing Piece
Engagement scores should decay over time. A demo request from 6 months ago isn't as valuable as one from last week.
Recommended Decay Schedule
| Time Since Action | Score Retention |
|---|---|
| 0-30 days | 100% |
| 31-60 days | 75% |
| 61-90 days | 50% |
| 90+ days | 25% |
Implementation: Most MAPs support score decay. In HubSpot, use a workflow that reduces engagement score by 25% every 30 days if no new activity.
MQL Threshold Calibration
Your MQL threshold should be calibrated against actual conversion data, not arbitrary numbers.
Calibration Process
- Apply your scoring model retroactively to the last 6-12 months of leads
- Segment by score ranges (0-25, 26-50, 51-75, 76-100)
- Calculate conversion rate to SQL and closed-won for each segment
- Set MQL threshold where conversion rate becomes acceptable to sales