Lead scoring assigns a value to each contact based on how likely it is to close. Linked to ads, it shows which campaigns, terms and creatives bring higher-value leads, and lets you teach platforms to look for more of them.
The essentials in a definition
Lead scoring assigns points to each contact based on form and behaviour data, to prioritise who deserves an immediate reply. Linked to ads, it shows which campaigns bring high-scoring leads and lets you send that signal to platforms, with well-defined conversions with value, events and key events. It should be simple, transparent and reviewed against real data.
What it is and how it works
Each lead receives points according to its characteristics (sector, size, role, budget, urgency) and behaviour (pages visited, response to emails, meeting requests). The sum indicates the degree of readiness to buy.
It can be done by hand, with simple rules, or with CRM and automation tools.
Linking scoring to ads
Store the source of each lead (campaign, ad, term) and its score. Analyse the average score and the percentage of hot leads per campaign. Compare cost per hot lead, not per total lead. That way you see where the value is.
Use the results to redistribute budget and improve messages.
Lead scoring linked to ads
- Define simple criteria and points
- Store the source of each lead
- Compare cost per hot lead by campaign
- Send hot leads to the platforms
Teaching platforms
Send platforms, as a conversion, the leads that pass a score threshold (for example hot lead) or values proportional to the score. That way the algorithm learns to look for people similar to the highest-value ones, not just anyone who fills in the form.
Ensure consent and good data quality.
Not all leads are worth the same; scoring says how much each one is worth.
Cautions
Review criteria against real closings: what looks good in theory may fail in practice. Avoid complicating: a few clear rules beat dozens. Avoid the score excluding valuable profiles through bias in the criteria.
Update regularly and document the rules.
Decision table
| Criterion | Example points | Source |
|---|---|---|
| Profile | Suitable size, sector and role | Form and public data |
| Intent | Requests a quote, demo or prices | Action taken |
| Behaviour | Repeat visits, pricing pages | Analytics with consent |
| Urgency | Short deadline and clear need | Qualifying question |
Common mistakes to avoid
- Opaque scoring. If nobody understands why, nobody trusts it.
- Criteria never validated. Check that a high score leads to more sales.
- Using data without a legal basis. Comply with data-protection rules.
How to measure the result
Validate quarterly the relationship between score and sale, and adjust weights; track response time by band.
Limits and exceptions
Scoring can bias decisions and exclude valid customers, and predictive models require ethical and legal care, especially with automated decisions. Incomplete or outdated data reduce reliability. Keep human oversight, review often and avoid criteria that may be discriminatory.
Illustrative example
Illustrative example: a company assigns 0 to 100 points based on five criteria and defines that leads above 70 get contact within 15 minutes. Comparing with sales, it sees leads above 70 convert six times more. It sends the platforms the high-score lead signal and sees cost per qualified lead fall.
In practice: checkpoints
Start with simple rules and a few criteria, and only later move to more complex models, when there are enough outcomes to train them. Share the score and its reasons with sales so they trust it, and compare conversion by score band to confirm the ranking makes sense. Check regularly that the score still predicts results after changes in campaigns, offers or markets, and keep human review for unusual cases. Document the criteria, the data used and who can change them, so the model does not become a black box that nobody can explain or correct.
Frequently asked questions
Do I need an expensive tool?
No. Simple rules in a spreadsheet or the CRM already help.
How many criteria should I use?
A few clear ones, validated with real closings.
How do I use scoring in ads?
By comparing campaigns by cost per hot lead and sending them to the platforms as a conversion.
Does scoring need AI?
Not necessarily; simple, transparent rules usually suffice, and predictive models require more data.
What data can I use?
Form data and on-site behaviour with consent, minimising and informing users.
A simple, validated score prioritises leads most likely to sell and, linked to ads, teaches platforms to look for similar contacts.
Sources and recommended reading
- Google Ads API: conversion management
- Analytics Help: GA4 recommended events
- Analytics Help: key events
Last editorial review: October 2026. Informational content; always confirm current conditions in the official sources before deciding.