B2B Lead Scoring: How to Prioritize the Right Leads
How B2B lead scoring works: models, criteria, tools, and how to make sure your sales team only talks to qualified leads.
Why Lead Scoring Is Decisive
Without lead scoring, your sales team treats all leads the same. The problem: SDRs spend just as much time on a freelancer who downloaded your whitepaper as on a VP Sales at an ICP company who visited your pricing page 3x.
With lead scoring:
- The top 20% of leads get immediate sales attention
- Mid-tier leads go into nurture sequences
- Unqualified leads get filtered out
- SDR productivity increases by 30-50%
The Scoring Model
Demographic Scores (Fit)
| Criterion | Score |
|---|---|
| Company size fits the ICP (50-500 employees) | +20 |
| Industry fits (SaaS, tech, B2B) | +15 |
| Role is a decision-maker (VP, C-level, Head of) | +20 |
| DACH region | +10 |
| Decision-maker in the right department | +10 |
| Max fit score | 75 |
Negative Scores (Disqualification)
| Criterion | Score |
|---|---|
| Freelancer / sole proprietor | -30 |
| Student / academic | -50 |
| Competitor | -100 |
| Wrong industry | -20 |
| Wrong country | -15 |
| Opt-out / unsubscribe | -100 |
Behavioral Scores (Engagement)
| Action | Score |
|---|---|
| Visited pricing page | +25 |
| Read case study | +15 |
| Requested demo | +40 |
| Used chatbot | +10 |
| Opened email | +3 |
| Clicked email link | +8 |
| Read blog post | +5 |
| Downloaded whitepaper | +10 |
| Attended webinar | +20 |
| Replied to outbound email | +30 |
Thresholds
| Score | Category | Action |
|---|---|---|
| 80+ | Hot lead (SQL) | Immediately to sales, contact within 1h |
| 50-79 | Warm lead (MQL) | Sales contacts within 24h |
| 25-49 | Nurture | Into email sequence, keep delivering content |
| <25 | Cold | No sales contact, passive nurturing |
Lead Scoring with AI
Predictive Lead Scoring
Instead of assigning points manually → AI analyzes historical deals and identifies patterns:
Input: 100 won deals + 200 lost deals AI analyzes: which combination of company size, industry, role, engagement, and timing correlates with winning? Output: automatic score per lead based on win probability.
Implementation with n8n + Claude
New lead in HubSpot
→ n8n fetches: company data, engagement history, enrichment
→ Claude API evaluates:
"Based on this data and our ICP:
Score this lead 1-100.
Justify it in one sentence."
→ Score → HubSpot property
→ If >70: Slack alert to SDRCommon Mistakes
- Model too complex: 50 criteria with fine-grained gradations → nobody understands it. Start with 10-15 criteria.
- Scores never adjusted: the model is created once and never validated. Check quarterly: do highly-scored leads really convert better?
- Fit only, no engagement: a CEO at an ICP company who never visited your website is less valuable than a manager who viewed the pricing page 5x.
- Marketing and sales not aligned: marketing defines the MQL, sales ignores the scores. Solution: shared definition and regular feedback.
Conclusion
Lead scoring is the lever that transforms your sales team from “treat all leads the same” to “the right leads at the right time.” Start simple (10 criteria, 3 thresholds), measure whether highly-scored leads convert better, and refine the model over time. With AI-based scoring, the model automatically gets better the more deals you close.
Common questions
What is lead scoring?
Lead scoring automatically evaluates leads based on criteria such as ICP fit (company size, industry, role) and engagement (website visits, email opens, content downloads). Each lead gets a score — the higher, the more sales-ready. Above a defined threshold, the lead is handed over to sales.
Which criteria belong in a lead scoring model?
Demographic: company size, industry, region, role, decision-making authority. Behavior: website visits (especially the pricing page), email engagement, content downloads, demo requests, chatbot interaction. Negative scores: wrong industry, freelancer, competitor, opt-out.
What is the difference between MQL and SQL?
MQL (Marketing Qualified Lead): has shown enough engagement to be handed over from marketing to sales. SQL (Sales Qualified Lead): has been contacted by sales and confirmed as a real opportunity (budget, authority, need, timeline). Not every MQL becomes an SQL — typical conversion: 30-50%.
Which tools do I use for lead scoring?
HubSpot (built in, from the Professional plan), Salesforce (Einstein Lead Scoring), custom via n8n + Claude API (the most flexible setup), Clay (score based on enrichment data). For SMEs: HubSpot or a custom n8n workflow. For enterprise: Salesforce Einstein or 6sense.