How does lead scoring work? Your CRM assigns points to each lead for fit attributes (industry, role, budget) and engagement actions (demo requests, pricing page visits), subtracts points for negative signals, and reduces old activity over time. When the total crosses a threshold you set from past conversion data, the lead is routed to sales.
If you're new to the concept, start with what is lead scoring. This guide is the mechanics: models, points, a worked example, and how to set a threshold that isn't a guess.
Key takeaways
- Every lead score is built from positive fit points, positive engagement points and negative points.
- Group caps stop one type of activity from inflating a score.
- Decay keeps scores current by shrinking old actions.
- Set your sales threshold by comparing conversion rates across score bands, not by picking a round number.
- Rule-based models work with little data; predictive and AI models need history.
The three lead scoring models
| Model | How it works | Data needed | Best for |
|---|---|---|---|
| Rule-based (manual) | You set points for each attribute and action | None beyond your CRM fields | New or small teams, clear ideal customer |
| Predictive | A machine learning model learns which traits preceded past conversions | Hundreds to thousands of past leads with outcomes | Teams with steady volume and clean history |
| AI (agentic) | Adds language models that read emails, chats and call notes to infer intent | CRM history plus conversation data | Teams with lots of unstructured conversations |
Salesforce, for instance, says its Einstein Lead Scoring needs at least 1,000 leads created in the last 200 days and at least 120 conversions, otherwise it falls back to a global model. We cover this in depth in predictive lead scoring and AI lead scoring. The rest of this guide focuses on rule-based scoring, which is where most teams start.
How lead scoring points work
Fit points
Fit points reward leads who look like your best customers. Pull this from closed-won deals: which industries, company sizes, roles and locations appear most?
Engagement points
Engagement points reward buying signals. High-intent actions (demo request, pricing page, replying to a quote) earn far more than low-intent ones (reading a blog post).
Negative points
Negative points remove noise: competitors, students, job applicants, unsubscribes, out-of-area enquiries.
Group caps and decay
HubSpot's lead scoring tool uses an overall score limit plus limits for each group of criteria, so 30 blog visits can't outscore one demo request. It also supports decay: its own example shows a 10-point form submission with 50% monthly decay dropping to 5 points after 30 days.
A lead scoring points table you can copy
This example is for a B2B services firm selling to Indian SMBs. Adjust the criteria to your business.
| Criteria | Type | Points |
|---|---|---|
| Owner, founder, director or department head | Fit | +20 |
| Company has 20-500 employees | Fit | +15 |
| Located in a city you serve | Fit | +10 |
| Requested a demo or quote | Engagement | +30 |
| Visited pricing page (last 14 days) | Engagement | +15 |
| Replied to email or WhatsApp message | Engagement | +15 |
| Clicked a link in a nurture email | Engagement | +5 |
| Attended a webinar | Engagement | +10 |
| Gmail/Yahoo address and no company given | Negative | -10 |
| Job seeker or student | Negative | -40 |
| Unsubscribed or asked not to be contacted | Negative | -50 |
| No activity for 60 days | Negative | -15 |
Notice there are no points for email opens. Apple Mail Privacy Protection preloads tracking pixels, which Mailchimp explains makes opens show as "opened" regardless of what the contact did.
Worked example: scoring three leads
Lead A: Priya, operations head at a 120-person logistics firm in Chennai. Requested a quote and visited pricing. Score: 20 + 15 + 10 + 30 + 15 = 90.
Lead B: Rahul, a student who downloaded two guides and clicked five emails. Score: 5 x 5 = 25 engagement points, minus 40 = -15, floored at 0.
Lead C: Anita, founder of a 10-person agency in Pune. Visited pricing and replied to a WhatsApp message. Company size is outside the band, so: 20 + 10 + 15 + 15 = 60.
Lead A goes straight to a rep. Lead C is borderline. Lead B is excluded, despite the most clicks, which is exactly what negative scoring is for.
How to set your sales threshold from data
Most teams pick 50 or 70 because it sounds right. A better approach builds on the method Salesforce describes: start from your baseline conversion rate and compare.
- Find your baseline. Say 400 leads last quarter produced 20 customers: a 5% lead-to-customer rate.
- Score those past leads with your new rules.
- Group them into bands and calculate the conversion rate of each.
| Score band | Leads | Customers | Conversion rate |
|---|---|---|---|
| 0-29 | 220 | 2 | 0.9% |
| 30-49 | 100 | 4 | 4% |
| 50-69 | 50 | 6 | 12% |
| 70-100 | 30 | 8 | 27% |
- Set the threshold where conversion clearly beats the baseline. Here, 50+ leads convert at more than twice the 5% average, so 50 is a sensible hand-off point. The 80 leads above it produce 14 of 20 customers.
The figures are illustrative, but the method works on any CRM export. If the bands don't separate, your rules aren't capturing what drives purchases. Check data quality first; our guide on CRM data quality explains why messy records break scoring.
What happens after a lead crosses the threshold
Scoring only matters if it triggers action. A typical setup:
- Score 70+: assign to a rep, create a call task due within the hour.
- Score 50-69: rep or AI agent qualifies with a short conversation using a framework like BANT (see how to qualify leads).
- Score under 50: stay in nurture; re-evaluate as engagement changes.
Under the hood, lead scoring evaluates each contact against predefined point thresholds. When an inbound inquiry accumulates sufficient points through demographic match and high-intent actions, the CRM automatically elevates its status to sales-ready. Automated alerts then notify account executives, ensuring prompt outreach while the prospect's buying interest is at its absolute peak.
Frequently asked questions
How are lead scores calculated?
In a rule-based model, the CRM adds points for each attribute or action that matches a rule, such as +20 for a target job title or +30 for a demo request, and subtracts points for negative signals. The total, often capped at 100, is the lead score. Predictive and AI models instead estimate a probability of conversion from historical data.
What is negative lead scoring?
Negative lead scoring subtracts points for signals that a lead is unlikely to buy: a competitor's email domain, a student or job-seeker, an unsubscribe, a location you don't serve, or long inactivity. It stops leads with lots of activity but no real buying potential from reaching sales and wasting reps' time.
How often should I update my lead scoring model?
Review it at least quarterly, and whenever you launch a new product, market or major campaign. Compare conversion rates for each score band. If high-scoring leads don't convert noticeably better than low-scoring ones, adjust the points or thresholds. Predictive models retrain automatically, but still need a human check.
What is lead score decay?
Score decay automatically reduces points from an action as it gets older, because a pricing page visit last week signals more intent than one six months ago. HubSpot, for example, lets you decay engagement points over set intervals, so a 10-point form submission with 50% monthly decay counts for 5 points after a month.
Conclusion: points are easy, calibration is the work
Lead scoring works by adding fit and engagement points, subtracting negatives, decaying old activity and routing leads above a threshold. The part most teams skip is calibration: checking that higher scores actually convert better, then adjusting.
Want lead scores that update automatically and trigger follow-up tasks? Try Autometa CRM free.
Related reading
- What Is Lead Qualification? Frameworks and Criteria
- Social Listening for Lead Generation: Find Buyers Before They Search
- How to Automate Follow-Up Emails (Gmail, CRM and AI)
- CRM in Social Media Marketing: A Practical Guide
Sources
- Overview of the lead scoring tool — HubSpot Knowledge Base, 2026
- What Is Lead Scoring? — Salesforce, 2026
- Considerations for Setting Up Einstein Lead Scoring — Salesforce Help, 2026
- Apple Mail Privacy Protection (MPP) FAQs — Mailchimp, 2026
