How to score leads in a CRM with enrichment data

Written by the Lusha content team. Lusha facts come from our documentation. HubSpot and Salesforce facts come from their own help pages, read on November 15, 2026, and are linked in Sources. The live example was run the same day.

To score leads in a CRM, you pick the fields that describe a good customer, give each one points, add points for recent activity, and set a threshold that tells a rep what to do next. HubSpot and Salesforce both do the arithmetic for you. What they cannot do is score a field that is empty or out of date.

This guide walks through the setup in six steps, shows where it lives in HubSpot and Salesforce, and ends with 16 accounts scored live for fit and timing so you can see what the numbers look like before you build.

Key takeaways

• A lead score has two halves. Fit asks whether this is the kind of company and person you sell to. Timing asks whether something is happening there now. Score both and keep them separate.
• Enrich before you score. A rule on industry or company size does nothing for a record where that field is blank.
• Set thresholds from your own numbers. In our live run the accounts that fit scored between 0.12 and 0.25, not 0.7 or above. A round number picked in advance would have missed all of them.
• Timing without fit misleads. Four of the six highest timing scores in our run belonged to accounts with a fit score of 0.
• Every tier needs an action. A score that does not change who calls, or when, is a number on a record.

What a CRM lead score is made of

Inbound teams lean on engagement because leads arrive with activity attached. Outbound teams have no engagement to score yet, so fit and timing carry the whole score. That is why lead scoring for outbound sales teams depends on enrichment more than on marketing activity.

How to score leads in a CRM in six steps

1. Define fit from the deals you won

Pull your closed won deals from the last 12 months and look for what they share: industry, employee range, country, and the title of the person who signed. Four or five fields is enough to start. Write them down as rules a stranger could apply.

2. Fill the fields the score will read

Check how many lead records have each of those fields filled in. If a third of your leads have no industry, a third of your leads cannot earn industry points. Enrich first. Lusha connects to HubSpot and Salesforce with field mapping, so company and contact fields land in the CRM fields your score reads. For a one time cleanup, run enrich and ICP score a lead list before it hits the CRM.

3. Give each field points

Weight the fields by how strongly they separate won deals from lost ones. A starting point for a team that sells to B2B software companies:

Example weights for illustration. They are not Lusha output. Set your own from your closed won deals.

4. Add timing

Fit changes slowly. Timing changes every week. Add a second score for what is happening at the account: a hiring surge, headcount growth, rising IT spend, a contact who moved into a new role. Lusha returns this as a Signal Score for any list of companies or contacts you bring, with the signals behind each score listed next to it. To rank leads by buying intent, sort by this score inside the accounts that already passed on fit.

5. Set thresholds from the spread of your scores

Score a few hundred existing records before you choose a cutoff. Look at where your won deals sit in the spread, then draw the lines. Three tiers are enough: work now, keep warm, leave.

6. Attach an action to every tier

Tier A goes to a rep with a task due today. Tier B goes into a sequence. Tier C stays out of the queue. Then review once a quarter: if Tier A does not convert better than Tier B, change the weights. To check the handoff, run grade the quality of inbound leads before they hit the AE.

How to set up lead scoring in HubSpot

HubSpot has a lead scoring tool under Marketing. Its documentation describes three kinds of score:

• Fit scores read property values such as job title, company size or annual revenue.
• Engagement scores read actions such as form submissions, email clicks and meetings.
• Combined scores use both and also show the fit and engagement values on their own.

You build a score from groups of rules. Each rule adds or subtracts points, each group can have its own limit, and the score has an overall limit. Engagement points can decay over time, so an old click counts for less. The result is stored as a score property that you can use in segments, workflows and reports. Which score types you can create depends on your HubSpot subscription, so check the HubSpot help page for yours.

The fit score is where enrichment does its work. Each property rule needs a value to read, so map industry, employee count, country, seniority and department from Lusha into the HubSpot properties your rules use.

How to set up lead scoring in Salesforce

Salesforce gives you two routes.

• Einstein Lead Scoring. A model scores each lead from 1 to 99 by how closely it resembles the leads you converted. The Salesforce implementation guide lists at least 1,000 leads created in the last 200 days, with at least 120 of them converted, to build a model on your own data. With less data it uses a global model. Setup asks which conversion milestone you use and which lead fields to leave out, and leads are scored again about every 10 days. Check Salesforce Help for the current requirements and editions.
• Rule based scoring. Add custom fields to the Lead object, then use a Flow to set a score when the record is created or updated. The Campus template score and route leads by ICP fit in Salesforce has the fields, the Flow and the routing.

Both routes read lead fields. Einstein learns from them and rules test them. Lusha CRM Data Enrichment updates Salesforce leads, contacts and accounts through field mapping, which keeps those fields filled as records come in.

A live example: 16 accounts scored for fit and timing

We took a list of company domains and scored it twice through the Lusha connector in Claude: once for ICP Fit Score against a test Target Profile, and once for Signal Score. The profile was B2B software companies with 51 to 1,000 employees in the United States and United Kingdom. The table is sorted the way a rep would work it.

Accounts scored live via Lusha connector, November 15, 2026. Company names are masked. Scores are shown as the connector returned them, on a scale of 0 to 1, rounded to two places. Tier rule used here: A is a fit score above 0 with a signal score of 0.5 or more, B is a fit score above 0 with a lower signal score, C is a fit score of 0. Fit reflects the test Target Profile, so the same list scores differently for a different seller.

What the run showed:

• The top timing score was the wrong account. Account I had every signal firing and a signal score of 1.00. It is a hospitality company, and its fit score was 0. Of the six highest signal scores on the list, four had a fit score of 0.
• Fit scores sat in a narrow band. The eight accounts that fit scored between 0.12 and 0.25. Eight scored 0, including the three that are not software companies. The profile treats company size as a must have, so a software company outside 51 to 1,000 employees also scores 0. The working threshold for this list was “above 0”, which is why step 5 says to look at the spread first.
• Timing sorted the accounts that fit. Seven of the eight had a signal score of 0.5 or more. One had a single signal and dropped to Tier B.
• The reasons came with the scores. Each signal score listed the signals behind it, which gives the rep an opening line as well as a rank.
• The cost was visible. The two scores on these 16 accounts used 32 credits. One more domain was not found and was not charged.

To bring scores like these into your CRM, add two number fields to the account or lead, one for fit and one for timing, and import the scored list into them. Lusha documentation describes importing CRM records into a Workspace table, enriching them and adding signals. It does not state that these scores are written to CRM fields for you, so plan the import as its own step.

Rule based or predictive: which to use

• Start with rules if you have fewer than a few hundred converted leads, a new product, or a team that needs to see why a lead ranked where it did.
• Add a predictive score once you have enough conversion history for a model to learn from. Use it to order the leads that already passed your fit rules.
• Keep both visible. A rep trusts a score when the reason is on the record: the fields that matched and the signals that fired.

CRM lead scoring for small businesses

A small team does not need a model. Use one fit score with four fields, one threshold, and one rule: anything above the line gets a call the same day. Check the fields are filled before you turn it on, and review the threshold after the first 50 leads. Add timing when you have more leads above the line than you can call.

Five mistakes that break a lead score

1. Scoring blank fields. Leads with missing data fall to the bottom for the wrong reason.
2. One number for fit and timing. A busy account that will never buy outranks a quiet one that will.
3. Thresholds picked before looking at the data. The cutoff has to come from the spread of your own scores.
4. No decay. A click from last year should not count like a click from last week.
5. No action per tier. If nothing changes for the rep, nobody reads the score.

FAQ

How do you score leads in a CRM?

Choose the fields that describe a good customer, give each one points, add points for engagement and for recent activity at the account, and set a threshold for each tier. HubSpot does this with fit, engagement and combined scores. Salesforce does it with Einstein Lead Scoring or with custom fields and a Flow.

What data do you need for lead scoring?

Company fields such as industry, size and country, contact fields such as title, seniority and department, engagement from your CRM and marketing tools, and buying signals such as hiring or a job change. The company and contact fields usually need enrichment, because forms collect few of them.

How do you rank leads by buying intent?

First remove the accounts that do not fit. Then sort the rest by a timing score built from buying signals. In Lusha that is Signal Score, which counts the active signals on a company or contact and lists them.

How does lead scoring work for outbound sales teams?

Outbound lists have no engagement yet, so the score rests on fit and timing. Score the list for fit against your ideal customer profile, score it for signals, and work the accounts that pass both first.

What is the difference between ICP Fit Score and Signal Score?

ICP Fit Score measures how closely a company or contact matches the Target Profile you set in Lusha. Signal Score measures how much buying activity is happening there now. Lusha combines the two into one Lead Score in its recommendations.

How often should a lead score update?

Fit can be rechecked when a record changes. Timing should refresh at least weekly, because signals expire. Review the weights and thresholds once a quarter against what converted.

Related: How sales lead scoring works, Lead scoring with Lusha, ICP scoring and the ICP Score and Route Skill.

Sources

All pages accessed November 15, 2026.

• Lusha docs: Signal Score and ICP Fit Score
• Lusha docs: CRM Data Enrichment
• Lusha docs: HubSpot integration
• Lusha docs: Salesforce integration
• Lusha docs: Importing from CRM into Workspace
• HubSpot Knowledge Base: Build lead scores
• HubSpot Knowledge Base: Understand the lead scoring tool
• Salesforce: Sales Cloud Einstein implementation guide (PDF)