HubSpot lead scoring ranks every contact by how likely it is to buy, using fit signals (who they are) and engagement signals (what they do). Manual rule-based scoring is a Professional-tier tool, and AI predictive scoring sits on Enterprise. For most Canadian manufacturers, wholesalers, and B2B teams, the better starting point is a manual model with separate fit and engagement scores, decay, and a written handoff threshold. Predictive scoring only pays off once you have years of clean closed-deal history, which low-volume, high-value sellers rarely do.
The Quick Answer
A lead score is a shortcut for a question every sales team asks daily: who should we call first? In HubSpot you can answer it with rules you write yourself (manual scoring) or with a model that learns from your past deals (predictive scoring). Manual is explainable and works with modest data. Predictive is powerful but hungry for clean history, and it costs more because it lives on a higher tier.
Why it matters: without a score, reps chase whoever emailed last or whoever is loudest. With a bad score, they learn to ignore it. A good one is simple enough that a rep can explain why a lead sits at the top of the list.
At-a-Glance: Manual vs Predictive Scoring
| Factor | Manual scoring | Predictive scoring |
|---|---|---|
| How the score is built | Rules you write: points added or removed per attribute and action | Machine learning on your historical closed deals |
| Explainability | Fully explainable: every point traces to a rule | Shows contributing signals, but you do not author the logic |
| Data needed | Your ideal customer profile and a few quarters of deal history | A large base of clean, attributed closed-won and closed-lost deals |
| Setup effort | A few weeks of design with sales, then quarterly tuning | Light setup, heavy data hygiene beforehand |
| Best fit | Most Canadian SMBs, low-volume high-value B2B sales | High-volume pipelines with years of reliable history |
| Typical HubSpot tier | Professional | Enterprise |
Why it matters: the data row decides most cases. A distributor closing a few hundred deals a year across thousands of contacts and a manufacturer closing twenty large accounts a year need very different tools, even though both want the same thing from a score.
1. Which HubSpot Tier Gives You Scoring
Lead scoring is not available on a Starter plan. Rule-based scoring arrives with the Professional tier, and AI predictive scoring is an Enterprise capability. HubSpot also rebuilt its scoring tooling in 2025, so older tutorials that refer to a single legacy score property may not match what you see in the portal today.
HubSpot changes edition names, seat rules, credit requirements, and pricing regularly. Confirm the current tier for the scoring feature you want, and the cost in Canadian dollars, on HubSpot's own pricing page before you commit budget.
Why it matters: the tier decision comes first. If you are on Starter, scoring is a reason to move up. If you are on Professional, you already own the tool that most SMBs need and the gap is design, not licensing.
2. Fit and Engagement: Build Two Scores
The most common design mistake is a single blended number. A contact who clicks every email but works at a company you cannot serve ends up with the same score as a perfect-fit buyer who has been quiet for a month. Those two leads need completely different handling.
Keep fit (industry, company size, role, region, buying channel) separate from engagement (pricing page visits, form fills, replies, meetings). Fit changes slowly. Engagement should decay. Hand a lead to sales only when both clear a threshold, and route high-fit, low-engagement accounts into targeted nurture or an account-based play instead.
Why it matters: two scores tell sales why a lead is ranked where it is. That explanation is what turns a score from a mystery number into a tool reps believe.
3. A Starter Scoring Model for B2B
Use this as a first draft to adjust with your sales team, not as a prescription. The point values are illustrative: set them so that a lead with strong fit and real buying behaviour clearly outranks everything else.
| Signal | Score type | Illustrative points |
|---|---|---|
| Industry matches your best customers (for example, a manufacturer or wholesaler) | Fit | +15 to +25 |
| Company size in your sweet spot | Fit | +10 to +20 |
| Buying-role job title (owner, operations, purchasing, finance) | Fit | +10 to +15 |
| Personal email address or non-business domain | Fit | -10 to -20 |
| Pricing, quote, or wholesale application page viewed | Engagement | +15 to +25 |
| Form submission or booked meeting | Engagement | +20 to +40 |
| Email clicks and repeat visits within 30 days | Engagement | +5 to +10 each |
| Competitor, student, or job-seeker indicators | Disqualifier | Set to a negative score that blocks handoff |
Why it matters:a first model does not need to be clever. It needs to be written down, explainable, and tested against last quarter's closed deals so everyone can see whether it would have ranked the right leads on top.
4. When Predictive Scoring Is Worth It
Predictive scoring learns from closed-won and closed-lost outcomes, so it is only as good as the history behind it. Third-party guidance commonly puts the practical floor at about two years of clean data and a few hundred closed deals properly attributed to HubSpot contacts. Below that, the model has little to learn from and tends to reward whatever was common rather than whatever was predictive.
That is a real constraint for the Atlantic Canadian manufacturers and wholesalers we work with, who often close a modest number of large accounts a year. If that sounds like you, invest in clean deal-to-contact attribution and a good manual model now, and revisit predictive scoring once the history exists. If your pipeline is high volume, predictive scoring can surface engagement and timing patterns a person would never write as a rule.
Why it matters: paying for an Enterprise capability that cannot learn from your data is the most expensive way to get a worse score than a spreadsheet of rules would give you.
5. Turn the Score Into Action
A score on its own changes nothing. Wire it into a workflow: when the fit and engagement thresholds are both met, update the lifecycle stage, assign an owner, create a follow-up task with a due date, and notify the rep. Below the threshold, keep the contact in nurture and let decay do its work. Sales and marketing should sign off on the threshold and the service-level expectation for follow-up, in writing.
Then close the loop. Every quarter, pull the deals that closed and look at what score each had a month earlier. Retune the rules that would have buried them. If your website forms, quote requests, and orders are not feeding HubSpot cleanly, fix that first, because a score built on missing data will confidently rank the wrong leads.
Why it matters: scoring is a feedback system, not a one-time setup. Teams that review it quarterly keep a model reps trust. Teams that set it once end up with a stale number everyone quietly ignores.
Which AtlanticWorks Guide Answers Your Question
Lead scoring touches several parts of HubSpot, and each has its own guide on this site. Use this table to land on the right one.
| If your problem is | Read |
|---|---|
| You want to know which hub and tier your automation needs overall | HubSpot Automation by Plan Tier in Canada |
| You want signals about which companies are researching you or your category | HubSpot Buyer Intent in Canada |
| You are designing the stages a scored lead moves through | How to Build a Sales Pipeline in Canada |
| You want the rep-facing hub where scores drive daily work | HubSpot Sales Hub Explained |
| You want AI agents that act on leads, not just rank them | HubSpot Prospecting Agent for Outbound AI |
| Your lead data is messy or out of sync with orders | Fix Ecommerce Order to CRM Sync Gaps |
Why it matters: scoring is one piece of a revenue system. Fixing it in isolation while pipeline stages, data sync, or tier choice are off just moves the problem somewhere else.
How AtlanticWorks Helps
AtlanticWorks implements HubSpot for Canadian manufacturers, wholesalers, and B2B teams. For scoring, that means defining your ideal customer profile with sales, building the fit and engagement model, wiring thresholds into routing workflows, cleaning the deal and contact data the model depends on, and setting up the quarterly review so the score keeps earning trust. We will also tell you plainly when your current tier is enough and when it is not.
Key Takeaways
- Lead scoring ranks who sales should call first. Manual rule-based scoring needs HubSpot Professional, and predictive AI scoring needs Enterprise, so confirm the tier on HubSpot's pricing page before you build.
- Build two scores, not one: a fit score for who the lead is and an engagement score for what they do. Blending them hides the difference between a great-fit account that went quiet and a poor-fit visitor who clicks everything.
- Predictive scoring needs a large base of clean, attributed closed deals. Many manufacturers and wholesalers sell too few high-value orders for it to learn from, so start with a manual model.
- Add decay and negative scores. Engagement points should fade, and competitors, students, and job seekers should never reach a rep.
- A score is only useful if a workflow acts on it. Agree the threshold, the lifecycle stage change, the owner assignment, and the follow-up task between sales and marketing in writing.
- Test the model against reality every quarter: check what score last quarter's closed-won deals had a month before closing, and retune the rules that would have ranked them low.
Frequently Asked Questions
What is lead scoring in HubSpot?
Lead scoring in HubSpot assigns a number to each contact, company, or deal based on signals that suggest how likely it is to become a customer. Those signals fall into two groups: fit, meaning who the lead is (industry, company size, role, location), and engagement, meaning what the lead does (visits, form fills, email clicks, meetings). Sales and marketing use the score to decide who gets follow-up first and when a lead is ready to hand over.
What HubSpot plan do you need for lead scoring?
Rule-based (manual) lead scoring is a Professional-tier feature, and AI-powered predictive scoring sits on the Enterprise tier. HubSpot adjusts editions, seat rules, and pricing regularly, so confirm which tier includes the scoring tool you want on HubSpot's own pricing page before you budget. A Starter plan does not include lead scoring.
What is the difference between manual and predictive lead scoring?
Manual scoring uses rules you write: add points for a job title, subtract points for a personal email address, add points for a pricing page visit. You control every input and can explain every score. Predictive scoring uses machine learning on your historical closed deals to estimate how likely a contact is to close, and it surfaces patterns a person would not write down. Predictive scoring needs a lot of clean historical data to work, which is why it suits larger teams and why a well-built manual model is the right start for most Canadian SMBs.
How much data do you need for HubSpot predictive lead scoring?
Predictive models learn from closed-won and closed-lost outcomes, so they need a meaningful history of deals that are properly attributed to contacts in HubSpot. Third-party guidance commonly cites roughly two years of clean data and a few hundred closed deals as a practical floor. If you sell a small number of high-value orders a year, as many manufacturers and wholesalers do, you will usually not have enough volume, and a manual fit-plus-engagement model will outperform a thin predictive one.
What should a B2B lead score include?
Build two separate scores rather than one blended number. A fit score covers firmographic and demographic criteria such as industry, company size, province or region, role, and whether the company buys through the channels you serve. An engagement score covers behaviour such as pricing or quote page visits, repeat sessions, form submissions, email replies, and meeting bookings, with decay so old activity counts for less. Keeping them separate lets sales see the difference between a great-fit company that has gone quiet and a poor-fit visitor who clicks everything.
How do you stop lead scores from going stale?
Add score decay so engagement points fall off after a set window, set negative scores for disqualifiers such as competitors, students, and job seekers, and review the model every quarter against actual closed deals. The most reliable test is to take last quarter's closed-won deals and check what score they had a month before they closed. If the model would have ranked them low, the rules need to change.
How does lead scoring connect to routing and follow-up in HubSpot?
A score does nothing until a workflow acts on it. Common patterns are to set a lifecycle stage such as marketing qualified lead when the score crosses a threshold, assign the contact to a rep, create a follow-up task, and notify the owner. Below the threshold, contacts stay in nurture. The threshold and the handoff rules should be agreed by sales and marketing in writing, because a score nobody acts on is just a number in a column.
Related Resources
Which tier unlocks which automation before you budget
Signals that show which companies are researching you
The stages a scored lead moves through
The rep-facing hub where scores drive daily work
Not sure your leads are being ranked the right way?
AtlanticWorks reviews how leads are captured, scored, and handed to sales, then tells you whether a manual model, a tier change, or a data fix will move the needle. The free assessment is the place to start.
Start the Assessment