Lead Scoring: How to Build a Model Your Sales Team Trusts

Lead scoring is how B2B teams decide which prospects deserve a sales call today and which ones need more time. Without it, reps cherry-pick leads by gut feeling, marketing argues that its leads are being ignored, and good opportunities go cold in the CRM. This guide walks you through building a lead scoring model from scratch, setting the MQL to SQL threshold with your sales team, and running it in HubSpot or Salesforce.

Key takeaways

• Lead scoring combines two dimensions: fit (does this account match your ICP?) and engagement (is this person showing buying intent?).
• A simple model with 8 to 12 criteria plus negative scoring beats a complex one nobody trusts.
• Your MQL threshold should be agreed with sales and based on closed-won deals, not picked at random.
• Every score band must trigger a specific action: a call, a nurture sequence, or a pause.
• Recalibrate quarterly by comparing scores at handoff with the deals that actually closed.

What is lead scoring?

Lead scoring is a method of ranking prospects by how likely they are to become customers. Each attribute or action adds or subtracts points, and the total (often on a 0 to 100 scale) tells sales who to prioritize. The point is not to grade leads for the sake of it, but to tell reps exactly who to call now and who to nurture.

Explicit vs. implicit scoring

Explicit scoring, also called fit or firmographic scoring, measures how closely a contact matches your ideal customer profile: industry, company size, job title, region, tech stack. This data comes from form fields or enrichment tools.

Implicit scoring, also called behavioral or engagement scoring, measures interest: pricing page visits, content downloads, webinar attendance, email replies, demo requests. It changes every time the lead interacts with you.

Negative scoring and score decay

A good model also takes points away: personal email domains, students, competitors, unsubscribes, long periods of inactivity. Add score decay too, so a click from six months ago weighs less than a click from yesterday. Without negative scoring, scores only go up and the model becomes meaningless within a few quarters.

Lead scoring models: which one fits your team?

There are three main types of lead scoring model. The right choice depends mostly on your lead volume and how much clean deal history you have.

• Rules-based point model: your team defines the criteria and the points. Data needed: low, your won and lost deals are enough. Strength: transparent and fast to launch. Weakness: built on assumptions you must validate. Best for: companies starting out.
• Fit x engagement matrix: two separate scores combined in a grid (A1 to D4). Data needed: moderate. Strength: separates a great-fit account that is not engaged yet from an engaged visitor who will never buy. Weakness: takes longer to configure. Best for: multiple segments or long sales cycles.
• Predictive lead scoring: a machine learning model learns from past conversions (HubSpot, Salesforce Einstein, Microsoft Dynamics 365). Data needed: a solid history of qualified and disqualified leads. Strength: spots patterns humans miss. Weakness: black box effect, useless with thin data. Best for: high volume and a clean CRM.

For most B2B companies, our recommendation is to start with a fit x engagement matrix and test predictive scoring once you have enough clean data to feed it.

How to build a lead scoring model in 7 steps

This is the process we follow in RevOps engagements. It fits into a few working sessions with marketing and sales.

1. Start from closed-won deals

Export your last 30 to 50 closed-won opportunities and the same number of closed-lost. Compare their attributes: industry, headcount, contact role, lead source, pages viewed before converting. The differences between the two groups are your first criteria.

2. Define your ICP and buying committee

Document your ideal customer profile at the account level and the roles involved in the purchase: economic buyer, champion, end user. A VP of Operations and an intern at the same company should not get the same score.

3. Pick your fit criteria

Keep it to 4 to 6 explicit criteria. Beyond that, the model gets hard to read and data gaps pile up.

4. Map engagement signals

Group actions into three tiers: weak signals (email open, blog visit), medium signals (content download, webinar), strong signals (pricing page, demo request, positive reply). An explicit demo request should be enough on its own to route a lead to sales.

5. Assign point values

Give more weight to the criteria that truly separate won from lost deals. Keep the scale simple, for example 0 to 100 for fit and 0 to 100 for engagement.

6. Set thresholds and actions

Each score band must trigger something: an automated nurture track, an SDR call, a handoff to an account executive. A score with no action attached is just a number.

7. Measure and recalibrate

Every quarter, compare lead scores at handoff with what happened next. If high-scoring leads never close, one of your criteria is overweighted.

Lead scoring example for a B2B company

Here is a sample model for a SaaS or industrial company selling to businesses with 50 to 1,000 employees. Treat the points as a starting point and tune them against your own deals.

• Industry in your priority segment: +20
• Company size between 50 and 1,000 employees: +15
• Decision-maker title (VP, director, C-level, procurement): +20
• Influencer title (project lead, technical manager): +10
• Located in a territory your sales team covers: +10
• Visited the pricing page: +15
• Downloaded an in-depth asset (guide, case study): +10
• Attended a webinar: +10
• Requested a demo or a quote: route straight to sales as an SQL
• Personal email domain (Gmail, Yahoo): minus 15
• Competitor or student: minus 50
• No activity in 90 days: minus 20

With this model, a lead above 60 points with an A or B fit grade becomes an MQL. An MQL who accepts a discovery call becomes an SQL.

MQL vs. SQL: setting the handoff threshold

A Marketing Qualified Lead (MQL) is a contact marketing considers ready for sales follow-up. A Sales Qualified Lead (SQL) is a contact sales has confirmed as a real opportunity, usually after a qualifying conversation using a framework like BANT (budget, authority, need, timeline).

Your MQL threshold has to be signed off by both marketing and sales, or it will be challenged in the first pipeline review. Put it in a simple service level agreement:

• A written definition of MQL and SQL, including the minimum score.
• A maximum response time for sales to work an MQL (for example, one business day).
• Required rejection reasons in the CRM when a rep disqualifies a lead.
• A monthly review of rejected leads to tune the model.

Leads below the threshold are not lost. They move into a nurture program until their score climbs back up.

How lead scoring works in your CRM

Most modern CRMs and marketing automation platforms include native lead scoring. The tool matters less than the quality of the data feeding it.

HubSpot, Salesforce and other lead scoring software

• HubSpot: a built-in scoring tool that lets you create fit and engagement scores for contacts, companies and deals, depending on your subscription.
• Salesforce: rules-based scoring through automation tools, plus Einstein Lead Scoring for predictive models. Account Engagement (formerly Pardot) adds its own scoring and grading.
• Microsoft Dynamics 365 Sales: predictive scoring that, according to Microsoft's documentation, requires at least 40 qualified and 40 disqualified leads created in the past two years.
• Adobe Marketo Engage, ActiveCampaign and Pipedrive also offer scoring, from advanced to lightweight.

Lead scoring best practices and mistakes to avoid

• Do not launch scoring on a dirty database full of duplicates, empty fields and untracked sources.
• Do not stack 30 criteria without knowing which ones actually predict closed-won.
• Always include negative scoring and decay.
• Never set the threshold without sales in the room.
• Recalibrate after launch, then every quarter.

Useful lead scoring is a living system, connected to pipeline and reviewed on a schedule.

Lead scoring FAQ

How do you calculate a lead score?

Add the points for every fit criterion the lead matches and every engagement action they took, then subtract negative points. For example, a director (+20) at a 200-person company in your target industry (+15, +20) who visited your pricing page (+15) scores 70. Compare that total with your MQL threshold to decide the next step.

What does lead scoring mean?

Lead scoring means assigning a numerical value to each prospect based on who they are and what they do, so sales and marketing can rank them by likelihood to buy. It turns a long list of contacts into a prioritized queue.

Can you give me an example of lead scoring?

A common example: +20 for a decision-maker title, +15 for a company in your size range, +15 for a pricing page visit, minus 15 for a personal email address. A lead who crosses 60 points is passed to sales. The full sample model above shows a complete grid.

How does lead scoring work in a CRM?

The CRM stores a score property on each contact or company. Rules or a predictive model update that property automatically whenever data changes or the lead takes an action. Workflows then use the score to trigger tasks, assign owners, or enroll the lead in nurture sequences.

How often should you update your lead scoring model?

Quarterly is a good default. Also revisit it whenever you launch a new product, enter a new segment, or add a major lead source.

Put your lead scoring to work

Lead scoring that actually drives revenue rests on three things: clean data, a shared definition of an MQL, and automation that triggers the right action at the right moment. That is exactly what our RevOps consulting team builds: CRM audit, scoring model designed with your sales team, setup in HubSpot or Salesforce, and ongoing tuning.

Not sure where your lead qualification stands? Get your free action plan and our experts will review your funnel and tell you what to fix first.

Related guides

• Marketing automation: a practical guide for B2B teams
• B2B lead generation strategies that build pipeline
• Sales funnel: how to build one that converts