HubSpot lead scoring: how to set it up, step by step
A practical walkthrough of HubSpot's scoring tool, from a blank portal to a score your workflows act on, with an example model, the settings that matter, and the mistakes to avoid.
HubSpot lead scoring is the scoring tool built into HubSpot’s Marketing Hub and Sales Hub. It gives contacts, companies, and deals a score from rules you set: points for who a record is (fit), points for what it has done (engagement), or both together (combined). Since 2025 it has replaced the legacy HubSpot Score property, which stopped updating on August 31, 2025.
This guide walks through setting it up from a blank portal to a score your workflows can act on, with an example model you can copy. Every step follows HubSpot’s own documentation on building lead scores and how the scoring tool works. HubSpot updates this tool often, so if a menu has moved, trust your portal.
HubSpot lead scoring at a glance
| Question | Answer |
|---|---|
| Where is it? | Marketing, then Lead Scoring |
| Which plans? | Marketing Hub or Sales Hub, Professional or Enterprise |
| What can it score? | Contacts (needs Marketing Hub), companies (either hub), deals (needs Sales Hub) |
| Score types | Fit (properties), engagement (events), combined (both). Deal scores are always combined |
| How points work | Rules add or subtract points inside groups. Each group is capped at 100 points by default |
| Decay | On event rules, a percentage reduction every 1, 3, 6, or 12 months |
| Bands | High, medium, and low for fit and engagement; A1 to C3 for combined scores |
| AI scores | Marketing Hub Enterprise, with at least 50 contacts: 25 that converted and 25 that did not |
| When does it update? | It scores records retroactively when turned on, then updates as properties and activity change |
Before you open HubSpot: plan the model
The tool is quick to use. The slow part is deciding what to score, and that is easier on paper than in a builder. Before you start, write down:
- Three to six fit factors, such as company size, industry, job level, and the systems a company runs. Take them from what your won deals have in common. Your ideal customer profile is the source.
- A handful of engagement signals, such as pricing page visits, demo requests, meetings, webinar attendance, and email replies. Leave email opens out; privacy features inflate them.
- Disqualifiers that subtract, such as competitors, students, personal email domains, and existing customers.
- Where your cutoffs should sit, so you know what “high” means before HubSpot asks.
If you want a structured way to do this, our five-step framework for building an ICP scoring model covers the fit side.
Two pieces of housekeeping first. Check that the properties you plan to score are actually filled in on most records, because an empty property earns no points. And if your portal used the legacy HubSpot Score, search your workflows, lists, and reports for it. Anything still reading it is reading a frozen number.
How to set up lead scoring in HubSpot, step by step
Step 1: Create the score and choose the object
Go to Marketing, then Lead Scoring, and click Create score. Choose what you are scoring:
- Contacts to rank people. Needs Marketing Hub.
- Companies to rank accounts. Works with either hub.
- Deals to rank open opportunities. Needs Sales Hub.
Most teams start with a contact score for inbound leads, or a company score if they sell to buying groups.
Step 2: Choose the score type
| Type | Scores on | Use it when |
|---|---|---|
| Fit | Property values | You want to know who matches your best customers |
| Engagement | Events (activity) | You want to know who is active right now |
| Combined | Both | You want one record to carry both, plus the A1 to C3 grid |
A combined score is usually the right choice for leads. It stores the fit value, the engagement value, and the total as separate properties, so you keep both answers instead of one blended number.
Step 3: Build the fit groups
Fit rules live in property groups. Click Add property group, then Add property criteria for each rule: choose the property, the operator, and the values, then whether the rule adds or subtracts points, and how many.
A few settings matter here:
- Score together or individually. When a rule lists several values, “score together” gives the same points for any of them. “Score individually” lets each value earn different points, which is how you score company size bands.
- Associated records. A rule can read a property on an associated record, such as the company behind a contact. When several associated records match, you choose how to combine them: sum, average, minimum, or maximum.
- Negative points. Put disqualifiers in their own group with points that subtract, so they are easy to find later.
Step 4: Build the engagement groups
Engagement rules live in event groups. Click Add event group, choose the event type, such as page views, form submissions, or meetings, and filter it, for example to your pricing page. Then set:
- A timeframe, such as the last 30 days.
- A frequency, if a single event is not enough.
- Points, added or subtracted.
- Decay, by turning on Decay scores and choosing a percentage and a period. HubSpot decays in steps of 1, 3, 6, or 12 months.
- A limit, by changing Score every time to Limit to and entering a maximum, so one busy visitor cannot run up the score.
Step 5: Set the group limits
Each group is capped at 100 points by default. You can change the cap per group, or use the Score limit setting to widen the overall range, from -100 to 100 up to -10,000 to 10,000. Keep the caps proportional to how much each group should matter. If company size should count for 30% of fit, cap that group at 30.
Step 6: Set thresholds and name the properties
Open the Settings tab. Under Score value, set clear labels and descriptions for the score properties so people know what they are reading. Under Score thresholds, set where high, medium, and low begin for fit and engagement. For a combined score, those thresholds produce the A1 to C3 grid, where the letter is fit (A is the highest) and the number is engagement (1 is the highest).
Set thresholds from your own data, not round numbers. Step 8 shows you how.
Step 7: Choose which records to score
On the contacts, companies, or deals tab, choose Score all, or Score specific to limit the score to a segment, such as one region or one product line. Scoring a defined segment is a good way to run a separate model for a separate motion.
Step 8: Test, preview, and turn it on
Before you turn anything on:
- Click Collapse all and read the whole model at summary level.
- Click Test a contact (or company, or deal) and pick records you know well. The test shows which rules added or subtracted points.
- Click Preview distribution to see how scores spread across your records. In large portals the preview is based on a sample, so treat it as a guide.
- Adjust thresholds so the high band holds a group small enough for sales to work.
- Click Review and turn on, then Turn on.
HubSpot scores every eligible record retroactively when the score turns on, based on current and historical values, and keeps it updated as properties and activity change.
An example HubSpot lead scoring model
Here is a combined contact score for a B2B software company selling to mid-market operations teams. Adjust the values to your own wins. Points are HubSpot’s; each group’s cap is shown.
Fit: property groups
| Group (cap) | Rule | Points |
|---|---|---|
| Company size (30) | Associated company has 200 to 1,000 employees | +30 |
| 50 to 199 employees | +15 | |
| Industry (25) | Associated company is in a core industry | +25 |
| Adjacent industry | +10 | |
| Role and seniority (25) | Job level is Director, VP, or C-level in operations | +25 |
| Manager in operations | +12 | |
| Systems they run (20) | Company uses the system your product needs | +20 |
| Disqualifiers | Email domain is a competitor | -100 |
| Personal email domain | -15 | |
| Lifecycle stage is Customer (new-business score only) | -100 |
Engagement: event groups
| Group (cap) | Rule | Points | Decay |
|---|---|---|---|
| Buying intent (40) | Demo or contact form submitted, last 90 days | +40 | 50% every month |
| Pricing and comparison (30) | Viewed pricing or comparison page, last 30 days, limit 2 | +15 each | 50% every month |
| Conversations (20) | Replied to a sales email or booked a meeting, last 30 days | +20 | 50% every month |
| Learning (10) | Attended a webinar or clicked a marketing email, last 90 days, limit 2 | +5 each | 50% every 3 months |
Thresholds: fit high from 70, medium from 40; engagement high from 50, medium from 20. Then check them against the preview and your won deals, and move them until the A1 and A2 bands hold a list sales can actually work.
Plays by band:
| Band | Play |
|---|---|
| A1, A2 | Assign an owner and create a task to reach out within one business day |
| A3 | Right company, not active yet. Add to ads and nurture, and alert the owner when engagement rises |
| B1, B2 | Qualify fast with a short call or a self-serve path |
| B3, C3 | Nurture only |
| C1, C2 | Check the fit data, then route to self-serve |
A demo request should route straight to sales whatever the band. The band sets the order of work, not whether someone gets an answer.
Putting the score to work
A score only matters once something acts on it. Score properties work anywhere a property does:
- Workflows. Enroll contacts when the combined score enters A1 or A2, then rotate an owner, create a task, set the lifecycle stage, or notify the rep.
- Lists and segments. Build a “work this week” list from the top bands, and a nurture list from A3 and B3.
- Views. Add the fit and engagement scores as columns in the contacts and companies views reps use every day.
- Reports. Track how many records sit in each band, and, more importantly, how each band converts to meetings, opportunities, and customers.
- Record cards. The score history card on each record shows which events changed the score.
Using HubSpot’s AI scores
On Marketing Hub Enterprise, you can let HubSpot draft the score. Choose Create score with AI, pick a fit or engagement score, then choose the conversion you care about as a lifecycle stage change (for example, subscriber to sales qualified lead) and a timeframe to evaluate. HubSpot evaluates your contacts, which can take up to an hour, and proposes criteria and points that you can review and edit before turning the score on.
It needs at least 50 contacts in the sample, with 25 that converted and 25 that did not. Treat the result as a first draft: read every rule, remove anything you cannot explain to a rep, and test it the same way as a score you built by hand.
Common setup mistakes
| Mistake | What happens | Fix |
|---|---|---|
| Old workflows still read the legacy HubSpot Score | Routing and alerts quietly stop working | Repoint them to the new score properties |
| Scoring email opens | Engagement inflates for people who never read anything | Score clicks, replies, and meetings |
| No group caps | One repeated action dominates the score | Cap each group in proportion to its importance |
| No decay | Last year’s activity outranks this week’s | Add decay to every event rule |
| Round-number thresholds | The top band holds too many or too few records | Set thresholds from the distribution preview |
| Scoring empty properties | Good leads score low because the data is missing | Fill the field or lower its cap |
| One score for every motion | The model fits one motion and misfires on the others | Use separate scores on specific segments |
When native scoring is enough, and when it is not
For a single product sold through one motion, HubSpot’s tool is often all you need. Teams tend to outgrow it when they need different decay speeds for different signals, buying group coverage that a sum or average cannot show, more models than their plan allows, signals that live outside HubSpot, or a reason on every record that a rep can read at a glance. Our guide to HubSpot lead scoring limits walks through each of those, with a workaround for each.
Where BoostIdeal fits
BoostIdeal is a scoring engine that works inside HubSpot, for the point where native scoring runs out. You build the model and set the weights, with AI helping you analyze your data, and every score and its reasons write back into HubSpot properties, so your workflows, lists, and reports keep working the way this guide describes. It scores any record, including leads, contacts, accounts, and custom objects, can score one record by the data on another, and runs several models side by side so you can compare them before anything acts on a score. Scores recalculate hourly by default, and in near real time with webhook sync.
Frequently asked questions
How do I set up lead scoring in HubSpot?
Go to Marketing, then Lead Scoring, and click Create score. Choose the object and the score type, add property groups for fit and event groups for engagement, set group limits, decay, and thresholds, then test records, preview the distribution, and turn it on.
Which HubSpot plans include lead scoring?
The scoring tool is available on Marketing Hub and Sales Hub, Professional and Enterprise. Contact scores need Marketing Hub, deal scores need Sales Hub, and company scores work with either. AI scores need Marketing Hub Enterprise.
What happened to the HubSpot Score property?
HubSpot retired it in stages in 2025. From August 31, 2025, legacy scores stopped updating and kept their last value. Any workflow, list, or report still using it should be moved to a score from the new tool.
Does HubSpot lead scoring decay?
Yes. Event rules in engagement scores can decay by a percentage you choose every 1, 3, 6, or 12 months. You can also use a workflow to reset an engagement score after a period with no activity.
What do A1 to C3 mean in HubSpot?
They are the bands of a combined score. The letter is the fit band, where A is the highest fit, and the number is the engagement band, where 1 is the most engaged. A1 is a strong-fit, highly engaged record; C3 is the opposite.
Does HubSpot have predictive lead scoring?
HubSpot offers AI-built scores on Marketing Hub Enterprise. HubSpot evaluates contacts that did and did not reach a lifecycle stage you choose, and proposes criteria and points you can review and edit before turning the score on.
Can HubSpot score companies as well as contacts?
Yes. Company scores are available with Marketing Hub or Sales Hub, and a company score can read the activity of its associated contacts, combined by sum, average, minimum, or maximum.
How often does HubSpot recalculate lead scores?
HubSpot evaluates records retroactively when a score is turned on, then keeps the score updated as property values and activity change. Editing a score re-evaluates records against the new rules.
The takeaway
Setting up HubSpot lead scoring takes an afternoon. Getting it right takes a plan before you start and a check after. Decide your factors and signals first, cap and decay every group, set thresholds from your own distribution, and wire the score into the workflows and views your team already uses. Then check every quarter that the top bands convert better than the rest.