Fit
Sector, size, region, role and use case.
Gedrag
Relevant pages, content, responses and repeat interactions.
Intentie
A specific enquiry, timing, problem and desired next step.
Negative signals
Poor target fit, outdated activity, unsubscribing or missing consent.
Validation
Compare scores regularly with actual sales outcomes.
What a lead score does and does not mean
Lead scoring orders contacts by predefined signals. It does not mean someone will certainly buy. It helps identify where human attention is likely to matter most and which contacts would be better served by a different path.
A high score without an explanation is of little use. Sales needs to see which signals contributed, when they were collected and whether the underlying data are still current.
Combine fit, behaviour and intent
Fit describes whether an organisation and contact match the chosen target audience. Behaviour shows interaction, but depends on context: a general download means something different from repeated visits to a relevant solution page. Intent becomes visible when a recognisable problem, timing or specific enquiry emerges.
Use negative scoring too. A student, supplier, existing customer enquiry, inactive contact or explicit unsubscribe should not follow the same route as a new business opportunity. Let behavioural scores decay as activity becomes outdated.
- Give strong, rare signals more weight than general activity.
- Avoid awarding duplicate points for repeated behaviour of the same type.
- Make exclusion and score-decay rules visible.
- Use segments when different target audiences behave differently.
- Document each field’s source, meaning and recency.
Set thresholds together with sales
Do not start with an arbitrary threshold. Analyse past contacts for which sufficiently reliable outcome data are available. Look for combinations of signals linked to useful conversations, and discuss false positives and missed opportunities.
Link each threshold to an action: direct personal follow-up, further qualification, nurturing or no commercial action. The path should match the strength of the signal and the contact’s expectations.
Sources for this section: European Union · European Data Protection Board
Illustrative example: two contacts with the same score
A fictional software company has two contacts with the same total score. The first is a strong target fit and has visited several substantive pages, but has not asked a specific question. The second has weaker target fit but has explicitly asked about an implementation timeline.
Instead of one ranking, the team uses separate dimensions for fit and intent. The first contact receives relevant follow-up content; the second is qualified by a person. The example shows why the composition of a score matters more than the total alone.
Treat scoring as profiling, with clear governance
When personal data are used to evaluate characteristics or behaviour, transparency matters. Define the purpose and legal basis, collect no more data than needed, keep it accurate and limit access and retention.
Do not let a score lead to a decision with significant consequences without appropriate review. Human judgement is also practically necessary for B2B prioritisation: context, relationships and current needs rarely fit entirely into a model.
Sources for this section: European Union · European Data Protection Board · Belgian Data Protection Authority
What does this mean for your organisation?
Scoring is useful when relevant incoming leads exceed what sales can follow up carefully and reliable outcome feedback is available. Start with a transparent model using a few signals and check monthly which scores lead to which outcomes.
Do not start if CRM data are incomplete, definitions differ or marketing receives no feedback on sales outcomes. Improve the data flow and shared qualification first.
Frequently asked questions
How many criteria does a good scoring model need?
As few as needed to distinguish contacts meaningfully. A simple, explainable model is usually easier to validate than dozens of weakly supported signals.
Is website behaviour enough for lead scoring?
No. Behaviour without target fit and context can mislead. Combine it with profile, intent and negative signals.
How often should you validate the model?
Monitor outcomes continuously and schedule joint reviews. Reassess sooner if the target audience, offer, campaign or data collection changes.
Is lead scoring allowed under the GDPR?
That depends on the purpose, data, legal basis, transparency and consequences. Assess the specific processing and provide suitable human oversight.
Sources
The sources below support the indicated factual and regulatory passages. The practical decision frameworks are professional recommendations from DSC Solution.
- European Union — Regulation (EU) 2016/679 — General Data Protection Regulation27 April 2016Used for: principles of lawful processing, data minimisation, accuracy and automated decision-making.
- European Data Protection Board — Automated decision-making and profiling6 February 2018Used for: profiling, transparency and the limits of solely automated decisions.
- Belgian Data Protection Authority — Information brochure on artificial intelligence systems and the GDPRDecember 2024Used for: Belgian considerations for AI and personal data.
