Back to Blog
HubSpot Guides Lead Scoring

How to Set Up Lead Scoring in HubSpot (The Right Way)

Most HubSpot lead scoring setups reward the wrong behavior and get ignored by sales within 60 days. Here is how to build a lead scoring model that actually drives action.

July 25, 2026 10 min read

HubSpot Lead Scoring

A score that sales does not trust is not a score. It is noise.

Lead scoring is one of the most commonly implemented and most commonly abandoned features in HubSpot. The pattern is predictable: a marketing team sets up a scoring model, sales ignores it after 60 days because the leads it surfaces are not actually better than the ones they were already working, and the score field sits in the CRM as a number that nobody looks at.

The problem is rarely the feature. It is the setup. Lead scoring fails when it rewards activity that does not correlate with buying intent, when the thresholds are not calibrated to real deal data, or when sales and marketing did not agree on what a qualified lead looks like before the scoring model was built. This post covers how to build a HubSpot lead scoring model that sales actually uses.

Before You Build: The Two Questions You Must Answer First

No lead scoring model should be built until you have clear answers to these two questions:

  • What does a contact who closed look like? Pull your last 20 to 30 closed-won deals and look at the contacts involved. What was their job title? Company size? Industry? Which pages did they visit before they became a lead? What actions did they take? These patterns are the foundation of your scoring model.
  • What does a contact who wasted a sales rep's time look like? Pull your lost deals and contacts who were qualified but did not buy. What patterns separate them from your best customers? These are your negative score triggers.

If you cannot answer these questions with data because your CRM does not have the history, start by defining your ICP in writing and use that as the basis for scoring. You will calibrate with data after 90 days.

The Two Components of HubSpot Lead Scoring

An effective HubSpot lead scoring model has two dimensions: demographic fit and behavioral engagement.

Demographic Scoring (Fit)

Demographic scoring rewards contacts who match your ICP. It answers the question "Is this person the right type of contact for us?" and does not change based on what the contact does. Examples:

CriteriaPoints
Job title is VP, Director, or C-level+20
Company size 50 to 500 employees+15
Target industry match+10
Company is in target geography+5
Personal email domain (gmail, yahoo)-20
Job title contains "student" or "intern"-20
Company size under 10 employees-10

Behavioral Scoring (Engagement)

Behavioral scoring rewards contacts for actions that correlate with buying intent. It changes over time as the contact engages. Examples:

ActionPoints
Visited pricing page+25
Requested a demo or booked a call+50
Opened 3+ emails in the last 30 days+15
Downloaded a case study+20
Visited the site 5+ times in 30 days+15
Watched a product video+10
No email opens in 90 days-15
Unsubscribed from email-50

How to Configure Scoring in HubSpot

In HubSpot, lead scoring is set up under Contacts > Properties > HubSpot Score (or your custom score property). The setup interface lets you add positive and negative scoring criteria using the same filter logic you use for lists and workflows.

Key configuration decisions:

  • Use HubSpot Score or create a custom score property? Use HubSpot Score for your primary model. Create a custom score property if you need a separate scoring track (for example, a product-qualified lead score separate from your marketing-qualified lead score).
  • Point decay: HubSpot does not have native point decay (where behavioral points expire over time if a contact goes cold). To approximate decay, add negative score criteria for contacts who have not engaged in 60 or 90 days.
  • Score-triggered workflows: Connect your scoring model to a workflow that changes lifecycle stage to MQL when a contact reaches your threshold, creates a task for the assigned rep, and sends an internal notification. The score alone is not enough. The action that follows the score is what makes it useful.

Lead Scoring Setup

Want a lead scoring model that sales actually trusts?

We build lead scoring models based on your closed deal data, calibrate the thresholds, and wire them to the workflows that make sales follow up. Book a call to start.

Book a Free Call

Setting the MQL Threshold

The MQL threshold is the score at which a contact becomes a Marketing Qualified Lead and is handed to sales. Setting it correctly is the most critical calibration decision in the whole system.

The right approach: look at your last 20 to 30 closed deals. Score each of those contacts retroactively against your model (manually or with a spreadsheet). Find the score value that most of those contacts would have had when they first became active leads. That number, plus or minus 10 percent, is your starting threshold.

If you do not have enough historical data, start at 50 points and calibrate after 30 to 60 days. Track the conversion rate from MQL to SQL (sales-accepted) and from MQL to closed deal. If your MQL-to-SQL rate is below 20 percent, your threshold is too low and you are sending too many unqualified leads to sales. If it is above 60 percent but your pipeline is thin, your threshold may be too high.

Why Lead Scoring Fails (And How to Prevent It)

The most common failure modes:

  • Too many positive triggers, no negative ones. If every action adds points and nothing subtracts them, every contact eventually reaches the threshold. The score stops meaning anything because even cold contacts who downloaded one ebook three years ago look like MQLs.
  • Rewarding form fills instead of intent signals. Downloading a top-of-funnel guide is not the same as visiting your pricing page. If your scoring model gives the same points to both, it is conflating awareness with intent.
  • Not reviewing the model after 90 days. Lead scoring is not set-and-forget. Review it quarterly: are the contacts scoring highest actually the ones converting? Adjust point values and thresholds based on what the data shows.
  • Not getting sales alignment before launch. If sales was not involved in defining what a good lead looks like, they will not trust the score when it arrives. Run a kickoff session with sales leadership before you build the model. Show them the criteria and get explicit sign-off on the threshold.

A lead scoring model that sales trusts and acts on is one of the highest-leverage things a RevOps team can build. It makes marketing's work visible, it gives sales a prioritization system, and it creates a shared definition of quality that improves alignment between the two teams over time. But it only works if it is built from deal data, calibrated against reality, and reviewed regularly. Build it once the right way and it pays for itself in the first quarter.

Common Questions

Frequently Asked Questions

There is no universal threshold. The right number is the one where most of your historical closed deals had a score above it when sales first engaged. A starting range of 50 to 80 points is common for B2B teams. Calibrate up if sales is overwhelmed with low-quality leads above the threshold. Calibrate down if your pipeline is too thin.

Standard contact scoring is rule-based: you define criteria and assign point values. AI lead scoring (Sales Hub Enterprise) analyzes your historical closed deal data to identify which attributes actually correlate with revenue, then weights them automatically. AI scoring updates as your data changes and tends to surface different contacts than rule-based scoring, often with higher conversion rates.

Yes. Negative scoring is as important as positive scoring. Without it, every contact eventually reaches the threshold and the score stops meaning anything. Add negative triggers for known non-ICP indicators: personal email domains, job titles like intern or student, company sizes outside your target range, and email unsubscribes.

Build it right the first time

A lead scoring model sales actually uses starts with your deal data.

We build HubSpot lead scoring models based on your closed deal history, calibrate the thresholds, and wire them to the automation that drives follow-up. Book a call to get started.

Book a Free Call