What Is a Lead Scoring Model?
A lead scoring model assigns numerical values to leads based on a combination of who they are (firmographic and demographic attributes) and what they've done (behavioral signals). The total score represents the lead's estimated likelihood to convert to an SQL or closed-won deal.
The purpose is prioritization and routing: sales reps should spend their time on the highest-probability leads, and those leads should be routed to the most appropriate rep. Lead scoring provides the data to make both decisions systematically rather than by gut feel.
Why Most Lead Scoring Models Don't Work
The failure mode is consistent: a marketing ops person builds a scoring model based on assumptions about what should predict conversion. The model gets deployed. Nobody validates it against actual conversion data. Years pass. The model scores leads based on criteria that are completely uncorrelated with actual pipeline outcomes.
The result: sales doesn't trust the scores, routes leads by other signals anyway, and the scoring model becomes an unused number in the CRM.
Building a model that works requires starting from conversion data — not assumptions.
The Two Dimensions of Lead Scoring
Dimension 1: Fit (Firmographic/Demographic Scoring)
Fit scoring measures how closely a lead matches your ideal customer profile. It answers: "Is this the type of company and person who typically buys from us?"
Common fit scoring attributes for B2B SaaS:
- Company size (employee count or ARR range)
- Industry vertical
- Geography / target market
- Job title and seniority level
- Technology stack (do they use tools that indicate readiness for your product?)
- Funding stage (recent funding = growth mode = budget)
Dimension 2: Intent (Behavioral Scoring)
Intent scoring measures buying signals — the lead's engagement with your brand and product. It answers: "Is this person actively evaluating solutions right now?"
Common behavioral scoring signals:
- Demo request form fill (+50 points — highest intent signal)
- Pricing page visit (+25 points)
- Multiple product page visits in same session (+20 points)
- Email engagement: opens and clicks (+5 points each)
- Webinar attendance (+15 points)
- Content download (+10 points)
- Return website visit (+5 points)
Building Your Scoring Model: Step-by-Step
Step 1: Pull Your Closed-Won Data
Export your last 12–18 months of closed-won opportunities. For each deal, record the firmographic attributes of the company (size, industry, geography) and the job title of the champion/buyer. This is your positive training set — what your ideal customer actually looks like.
Step 2: Identify Negative Patterns
Export your lost or disqualified leads over the same period. Look for patterns: which company sizes, industries, or job titles consistently show up in unqualified or lost deals? These become negative scoring criteria — attributes that lower a lead's score.
Step 3: Assign Point Values Based on Conversion Correlation
For each attribute, estimate how strongly it correlates with conversion. Attributes where 80% of closed-won deals match a criterion get high positive scores. Attributes where 5% of closed-won deals match get negative or zero scores.
Start with this rough framework:
| Attribute | Points | Rationale |
|---|---|---|
| Company size: 50–500 employees | +20 | Core ICP range |
| Company size: 500–2,000 employees | +15 | Upper mid-market — still converts well |
| Company size: < 20 employees | -10 | Below minimum viable deal size |
| Job title: VP Sales, CRO, Head of RevOps | +25 | Primary buyer persona |
| Job title: Sales Manager, AE, SDR Manager | +15 | Champion / power user persona |
| Job title: Engineer, Developer | -15 | Not a buyer for this product |
| Industry: B2B SaaS, tech, fintech | +20 | Highest-converting verticals |
| Demo request form fill | +50 | Explicit buying intent |
| Pricing page visit | +25 | Commercial evaluation signal |
Step 4: Set the MQL Threshold
The MQL threshold is the score at which a lead is considered marketing-qualified and handed to sales. Setting it too low floods sales with unqualified leads. Setting it too high means qualified leads never reach sales.
The right threshold is data-driven: it's the score at which your historical conversion rate to SQL shows a meaningful inflection point. If leads scoring 60+ convert at 30% and leads scoring 40–59 convert at 8%, the threshold is around 60.
Step 5: Implement Score Decay
Behavioral scores should decay over time. A lead who visited your pricing page 18 months ago is not as high-intent as one who visited yesterday. Configure score decay to reduce behavioral points by 50% after 30 days and 100% after 90 days. Fit scores (firmographic) don't decay — a company's size doesn't change.
Step 6: Validate and Recalibrate Quarterly
Every 90 days, review: what is the SQL conversion rate by score band? Are high-scored leads converting at higher rates than low-scored leads? If not, your scoring weights are wrong. Adjust them based on actual conversion data, not assumptions.
How to Use Lead Scores for Routing
Lead score is a powerful routing signal. Here's how to use it:
- Score 80+: Route directly to an AE with an instant Slack notification. These are your highest-intent leads — they should never wait in a queue.
- Score 50–79: Route to an SDR for qualification first. They may be evaluating options but aren't ready for an AE conversation immediately.
- Score 20–49: Enter nurture sequence. Marketing automation, not sales, should work these until they score higher.
- Score < 20: Disqualify or ignore until behavioral signals increase the score.
For the full routing model that combines score with territory and account ownership, see our guide on MQL routing for B2B SaaS. For conversion benchmarks by routing path, see our MQL to SQL conversion rate guide.
Route leads by score, territory, and availability.
Lead Dispatcher combines your lead score with CRM data and rep availability to dispatch inbound meetings to exactly the right person, instantly.
Book a DemoFrequently Asked Questions
What is a lead scoring model?
A lead scoring model assigns numerical values to leads based on firmographic attributes (who they are) and behavioral signals (what they've done) to rank leads by their likelihood to convert. It's used for prioritization and routing decisions.
What are the two types of lead scoring?
Fit scoring (firmographic/demographic) assesses ICP match — company size, industry, job title. Intent scoring (behavioral) measures buying signals — pricing page visits, demo requests, email engagement. Effective models combine both dimensions.
What score should trigger MQL status?
Set your MQL threshold at the score where your historical data shows an inflection point in SQL conversion rate. If leads scoring 60+ convert at 30% and 40–59 convert at 8%, your threshold should be around 60.
How do you validate a lead scoring model?
Compare lead scores to SQL conversion rates over time. If high-scored leads convert at significantly higher rates than low-scored leads, the model has predictive validity. Recalibrate quarterly based on actual conversion data.
How does lead scoring affect lead routing?
Lead score directs leads to the right routing tier: high scores go directly to AEs for immediate engagement, mid-range scores go to SDRs for qualification, low scores enter nurture. This improves rep-to-lead match quality and conversion rates.