Article content and detailed guides remain in English. The selected language applies to controls and quick instructions.

Back to articles

Use cases

Behavioural Lead Scoring: Using Intent and Engagement Signals to Prioritise Prospects

By Email ExtractorPublished 9 min read
On this page

Demographic vs Behavioural Scoring

Traditional lead scoring assigns points based on who the lead is: their job title, company size, industry and location. Behavioural scoring adds a second dimension: what the lead does.

Scoring type Based on Answers the question Limitation
Demographic Who they are (title, company, industry) "Is this lead a good fit?" A perfect-fit lead who is not in-market will not convert now
Behavioural What they do (visits, clicks, downloads, engagement) "Is this lead interested now?" A highly engaged lead at the wrong company is still a bad fit
Combined Both fit and behaviour "Is this the right lead at the right time?" More complex to build and maintain

The most effective scoring systems combine both: demographic scoring qualifies the lead, and behavioural scoring indicates timing and urgency.

Behavioural Signals and What They Mean

First-party signals (from your own channels)

Signal Strength What it indicates
Pricing page visit Strong Evaluating cost; close to a decision
Demo request or free trial signup Very strong Active evaluation
Case study or testimonial page visit Strong Building a business case; may be comparing vendors
Multiple blog visits in a short period Medium Researching the problem area
Email opens (multiple) Weak Some interest; may be habitual opener
Email link clicks Medium Engaged with specific content
Content download (whitepaper, guide) Medium Interested in the topic
Webinar registration Medium Actively learning about the space
Webinar attendance Strong Invested time; more serious than registration alone
Return visit after a gap Medium Renewed interest; something triggered re-engagement
Contact page visit Strong Considering reaching out
Integration or API documentation visit Strong Evaluating technical fit
Comparison page visit (your product vs competitor) Very strong Actively comparing options
Unsubscribe Negative Not interested; remove from active scoring
Form abandonment Weak Some interest but not enough to complete the form

Email engagement signals

Signal Points to assign Notes
Open (first time) +1 Low value on its own; email open tracking is unreliable
Open (repeat of same email) +2 Re-reading suggests interest
Click on content link +5 Engaged with the content
Click on pricing/demo link +10 High-intent action
Reply +15 Active engagement
Forward +10 Sharing with colleagues; possible internal championing
Unsubscribe -20 Clear disinterest
Spam complaint -50 Remove from all scoring and outreach

Third-party intent signals

Signal Source What it indicates
Researching your product category Bombora, G2, TrustRadius intent data In-market for your type of solution
Visiting competitor websites Bombora, similar intent data providers Actively evaluating options
Searching for related keywords Google Ads keyword data, SEO tools Problem-aware and researching solutions
Published RFP or RFI Government procurement sites, RFP databases Active buying process
Job postings for related roles Job boards, LinkedIn Building a team; may need tools for the function
Technology change (removing or adding tools) BuiltWith, Wappalyzer technographic data Stack change creates integration needs
Review left on competitor product G2, Capterra, TrustRadius Using a competitor; may be dissatisfied
Social media discussion about the problem LinkedIn, Twitter/X, Reddit Problem-aware; engaging publicly
Conference attendance in your space Conference attendee lists, social media Investing time in the problem area

Building a Behavioural Scoring Model

Step 1: Define scoring tiers

Score range Classification Sales action
0-20 Cold Nurture with educational content
21-40 Warming Continue nurture; increase content frequency
41-60 Warm Marketing qualified lead (MQL); prioritise for targeted content
61-80 Hot Sales qualified lead (SQL); sales outreach
81-100 Very hot Immediate sales follow-up; likely in active evaluation

Step 2: Assign point values

Category Signal Points Decay
Website Pricing page visit +15 -5 per week
Website Blog post visit +3 -1 per week
Website Case study page visit +10 -3 per week
Website Multiple pages in one session (3+) +8 -2 per week
Website Return visit (within 7 days) +5 -2 per week
Email Open +1 -0.5 per week
Email Click +5 -2 per week
Email Reply +15 -3 per week
Content Download whitepaper/guide +8 -2 per week
Content Attend webinar +12 -3 per week
Event Visit booth at trade show +15 -3 per week
Event Attend your presentation +10 -3 per week
Third-party Intent data signal (researching category) +12 -4 per week
Third-party Competitor review posted +8 -2 per week
Negative Unsubscribe -20 No decay
Negative Spam complaint -50 No decay
Negative No activity for 30 days -15 One-time
Negative Bounced email -30 No decay

Step 3: Implement score decay

Behavioural scores should decrease over time because interest fades:

Decay approach How it works Advantage
Linear decay Subtract fixed points per time period Simple; predictable
Percentage decay Reduce score by a percentage each period High scores decay faster in absolute terms
Event-based decay Reduce score after a period of no activity Only penalises true inactivity
Threshold decay Drop score to a lower tier after inactivity Clear tier transitions

Example with linear decay:

A lead visits your pricing page (+15 points), then does nothing for 3 weeks. With -5 per week decay on that signal, the pricing page contribution drops to 0 after 3 weeks. If they then open an email (+1), their score is 1, not 16.

Why decay matters: Without decay, a lead who visited your pricing page six months ago and never returned would still appear "hot." Decay ensures the score reflects current interest.

Step 4: Set thresholds for action

Threshold Trigger Action
Score reaches 40 (MQL) Marketing handoff Add to targeted campaigns; increase personalisation
Score reaches 60 (SQL) Sales notification Sales rep reviews lead and initiates outreach
Score reaches 80 (hot) Priority alert Immediate personalised outreach from sales
Score drops below 20 after being above 40 Re-nurture Move back to educational content; reduce frequency
Score drops below 0 Suppress Remove from active campaigns
Negative signal (unsubscribe, spam) Immediate Remove from all scoring and outreach

Data Sources for Behavioural Scoring

Setting up tracking

Data source How to capture Tools
Website visits JavaScript tracking pixel, server logs Google Analytics, HubSpot, Segment
Email engagement Email platform tracking Mailchimp, HubSpot, ActiveCampaign
Content downloads Form submissions, gated content Marketing automation platform
CRM activity Sales rep logging interactions Salesforce, HubSpot CRM, Pipedrive
Third-party intent Data provider integration Bombora, G2, ZoomInfo
Social engagement Social listening tools, API access Hootsuite, Sprout Social, native APIs
Event attendance Registration systems, badge scans Event platforms, CRM integration

Connecting behavioural data to email addresses

The challenge with behavioural scoring is linking anonymous website visits to identified leads:

Identification method How it works When it triggers
Form submission Visitor fills in a form with their email Content download, demo request, newsletter signup
Email click tracking Visitor clicks a link in your email; cookie identifies them Any email link click
CRM record match IP or cookie matches a known CRM contact Returning website visitors
Account-level identification IP address mapped to a company (not individual) Any website visit from a corporate network
Login Visitor logs in to your product or portal Existing customers or trial users
Chat widget Visitor provides email in a chat conversation Support or sales chat interactions

For identified leads, extract and consolidate their contact information from various sources. If you have exported files from event registrations, CRM exports or form submissions, upload them to Email Extractor to deduplicate email addresses across sources before feeding them into your scoring system.

Scoring Model Maintenance

Regular calibration

Activity Frequency Purpose
Review score distribution Monthly Ensure scores are distributed meaningfully (not all bunched at one end)
Check conversion correlation Quarterly Verify that high-scoring leads actually convert at higher rates
Adjust point values Quarterly Increase points for signals that predict conversion; decrease for those that do not
Review decay rates Quarterly Ensure decay is fast enough to prevent stale scores
Add new signals As available Incorporate new data sources as they become available
Remove noisy signals As identified Drop signals that do not correlate with conversion
Re-calibrate thresholds Semi-annually Adjust MQL/SQL thresholds based on sales feedback and conversion data

Common scoring model problems

Problem Symptom Fix
Score inflation Too many leads are "hot" Increase decay rates; raise thresholds; reduce points for weak signals
Score stagnation Leads never reach MQL Lower thresholds; add more trackable signals; review if enough engagement channels exist
False positives High-scoring leads do not convert Review which signals correlate with conversion; remove those that do not
False negatives Converting leads had low scores Add signals you are not tracking; review if key pages lack tracking
Competitor research traffic Competitor employees score high from researching you Exclude known competitor domains from scoring
Bot traffic Automated visits inflate scores Filter bot traffic before scoring
Content-only leads Leads engage with educational content but never evaluate the product Differentiate between educational and commercial intent signals

Advanced Techniques

Account-level scoring

For B2B, scoring at the account level (multiple contacts at the same company) can be more useful than individual scoring:

Metric How to calculate What it indicates
Number of contacts engaged Count unique contacts from the same company who have interacted Breadth of interest within the organisation
Contact seniority mix Are decision makers (VP, C-suite) engaged, or only researchers? How far along the evaluation is
Total account score Sum of individual scores from the same company Overall account engagement level
Multi-threading depth Number of different departments engaging Cross-functional evaluation (stronger buying signal)
Score velocity How quickly the account score is increasing Urgency of evaluation

Predictive scoring

Traditional point-based scoring is rules-based. Predictive scoring uses historical data and machine learning:

Approach How it works When to use
Rules-based (manual) Assign points based on experience and hypotheses Starting out; small datasets; simple sales process
Regression model Statistical model trained on historical conversion data Medium-size dataset; want data-driven weights
Machine learning Algorithm learns patterns from all available data Large dataset; complex buying patterns; many signals
Hybrid Rules-based for known signals + ML for pattern discovery Best of both; interpretable core with ML enhancement

Extract emails

Explore tools

Verify emails

Check address validity before using your list.

ZeroBounce

Email Verification

Verifies email lists and provides tools for monitoring deliverability.

Useful when list cleaning and sender health belong in one workflow.

Explore ZeroBounce (opens in a new tab)