Salesforce lead scoring: four ways to score leads, from formula fields to Einstein
Salesforce has no single lead score. It has four ways to get one, and each needs different data. Here is how they work, what they require, and how to choose.
Salesforce lead scoring is any method of ranking leads inside Salesforce so reps know who to work first. There are four ways to do it today: a rules-based score you build with formula fields and Flow, Einstein Lead Scoring in Sales Cloud, the score and grade in Marketing Cloud Account Engagement (formerly Pardot), and third-party scoring tools that write a score into Salesforce.
The right one depends on your edition, your lead volume, and how much say you want over the logic. Everything here about Salesforce comes from Salesforce’s own help pages and Trailhead, linked where it is used. Packaging changes often, so check your edition before you plan around a feature.
Salesforce lead scoring at a glance
| Question | Answer |
|---|---|
| Is there one built-in lead score? | No. Salesforce has several scoring features in different products, plus the tools to build your own |
| Build it yourself | A formula field for fit, a record-triggered flow for points that change with activity |
| AI score in Sales Cloud | Einstein Lead Scoring, part of Sales Cloud Einstein |
| Einstein data minimum | 1,000 leads created and 120 converted in the last 200 days, per segment |
| Without enough data | Einstein uses a global model built on anonymous data from many Salesforce customers, if you score all leads together |
| Marketing automation scoring | Account Engagement score (activity) and grade (fit, F to A+) |
| AI score for behavior | Einstein Behavior Scoring in Account Engagement, 0 to 100 |
The four options side by side
| Rules you build (formula, Flow) | Einstein Lead Scoring | Account Engagement score and grade | Third-party tool | |
|---|---|---|---|---|
| Who sets the logic | You | Salesforce’s model, from your conversion history | You, starting from Salesforce’s default rules and profile | Depends on the vendor: you, their model, or both |
| Data needed | Filled-in lead fields | 1,000 leads and 120 conversions in 200 days for a custom model | Tracked prospect activity and prospect fields | Varies. Check with the vendor |
| Can reps see why | Only if you expose the parts on the page | Yes, the top contributing fields | Score: activity history. Grade: the profile criteria | Varies. Ask to see a scored record |
| Setup effort | Low for a formula, medium for Flow | Low to turn on, up to 48 hours to score | Medium: edit rules, build profiles, add automation | Varies |
| Best for | Teams that want full control, or lack conversion volume | High lead volume with clean conversion history | Marketing teams running Account Engagement | Signals that live outside Salesforce |
Option 1: a rules-based score with formula fields and Flow
This is the route with no extra license. You decide the criteria and the points, and Salesforce does the arithmetic.
Formula fields for fit. A formula field calculates its value from other fields and updates on its own when those fields change. Salesforce’s formula data types include Number, which is what a score needs. You create one in Object Manager under Fields and Relationships, as Trailhead’s formula fields unit shows. A formula works from field values, so it suits fit: industry, company size, title, country.
Flow for engagement. Activity is different, because points have to be added when something happens. A record-triggered flow runs when a lead is created or updated. Salesforce’s page on before-save record-triggered flows explains that a flow can update the triggering record before it is saved, and that you need an after-save flow to create or update related records. So a flow can write points into a number field on the lead when a form, a campaign response, or a status change arrives.
How to build a rules-based score in six steps
- Write the model down first. Pick three to five fit factors from what your won deals have in common, and a short list of actions that show buying interest. Our five-step framework for building an ICP scoring model covers the fit side.
- Check the data. Look at how many leads have each field filled in. A factor that is empty on most leads will sink good leads.
- Create a Fit Score formula field on the lead, with Number as the return type. Give each factor a point cap in proportion to how much it should matter.
- Create an Engagement Score number field, and a record-triggered flow that adds points when a qualifying action is recorded.
- Add both fields to the lead page and list views, next to the fields they read, so reps can see the inputs.
- Test on leads you know. Score 20 leads that converted and 20 that did not. If the two groups look the same, change the model before anyone routes on it.
A fit formula can be as simple as this sketch. Test it in a sandbox against your own field values:
CASE(Industry, "Manufacturing", 25, "Technology", 25, "Healthcare", 10, 0)
+ IF(AND(NumberOfEmployees >= 200, NumberOfEmployees <= 1000), 30,
IF(AND(NumberOfEmployees >= 50, NumberOfEmployees < 200), 15, 0))
+ IF(OR(CONTAINS(Title, "Director"), CONTAINS(Title, "VP")), 25, 0)
- IF(CONTAINS(Email, "@gmail.com"), 15, 0)The limits of this route are practical. Every rule is yours to maintain, long formulas get hard to read, and points for activity need a plan for removing them when the activity gets old.
Option 2: Einstein Lead Scoring
How it works
Einstein Lead Scoring is the AI score in Sales Cloud. Salesforce describes it this way: “Einstein analyzes your past leads to determine which current leads have the most in common with leads that have previously converted.” You do not write rules or set points. Trailhead adds that it analyzes historical field data on the lead object, including custom fields.
What it needs
Salesforce’s setup considerations list the minimums for a model built on your own data:
| Requirement | Detail |
|---|---|
| Leads | At least 1,000 leads created in the last 200 days |
| Conversions | At least 120 of those converted to an account and contact |
| If you count conversion as opportunity creation | At least 120 converted with an opportunity created at conversion |
| Segments | The minimums apply to each lead segment you set up |
| Encrypted fields | Not used in the analysis |
| New lead fields | Must be added by hand to each segment’s included fields |
If you score all leads together, without segments, and you do not have enough conversion data, Einstein falls back to a global model, which “uses anonymous data from many Salesforce customers.” That gets a score on the page, but it is a model of other companies’ leads, not yours.
Where the score and the reasons appear
The score shows in the Einstein Score component on lead detail pages, and the component shows which of the lead’s fields had the greatest influence. When the Einstein Score field is added to a list view, hovering over a score shows the top factors behind it. Field-level security applies, so a rep only sees the factors for fields they are allowed to read.
Setup, refresh, and editions
The setup steps are short: turn it on in Setup, choose the conversion milestone (account and contact, or opportunity creation), choose whether to score all leads together or in segments, and choose which lead fields to include. Salesforce says it can take up to 48 hours to add scores, and that Einstein reanalyzes lead data every 10 days and refreshes them.
Salesforce’s help pages list Einstein Lead Scoring with Sales Cloud Einstein, which is included in Performance and Unlimited Editions and available for an extra cost in Enterprise Edition. Check your edition and contract before you plan around it.
It suits a team with steady inbound volume and consistent lead conversion habits. It is weaker when reps convert leads inconsistently, when you sell to a new segment with no history, or when your most important fit data is not on the lead record.
Option 3: Account Engagement score and grade
Marketing Cloud Account Engagement, the product formerly called Pardot, gives each prospect two values. Salesforce’s glossary defines them plainly: a score “is based on activities (pages viewed, forms completed, and so on) that a prospect performs,” and grading “reflects a prospect’s persona, fit, or demographic.”
The score. Account Engagement ships with a default scoring system that you can edit. A few of the defaults:
| Activity | Default points |
|---|---|
| Form submission | +50 |
| Opportunity created | +50 |
| Video watched 75% or more | +25 |
| Page view | +1 |
| Email open | 0 |
| Form error | -5 |
| Opportunity lost | -100 |
Email opens score zero by default, which is a sound choice: Litmus reports that over 50% of email opens come from devices with Apple Mail Privacy Protection, so an open says little about interest. Scores can also be changed by automation rules and completion actions, and scoring categories keep a separate score per product or business unit. Decay is something you set up yourself: Trailhead’s scoring unit describes it as reducing a prospect’s score for inactivity, typically with automation rules.
The grade. Per Trailhead’s grading unit, grades run from F to A+, every prospect starts at D, and the default profile grades against five criteria: company size, industry, location, job title, and department. Matches move the grade up or down in fractions of a letter, such as one third or two thirds.
In Salesforce. Both values sync to read-only fields on the lead and contact, named Account Engagement Score and Account Engagement Grade in Salesforce’s default field mapping. Account Engagement is the source of truth for both.
Einstein Behavior Scoring. This is the AI counterpart to the activity score. Salesforce’s page on how Behavior Scoring works says it assigns a score from 0 through 100 based on engagement patterns, and that tailored insights need six months of engagement data for connected prospects and at least 20 prospects linked to opportunities. It appears in a Behavior Scoring component on lead and contact pages. The setup page lists it for Account Engagement Advanced and Premium Editions with Salesforce Enterprise, Performance, or Unlimited.
Option 4: third-party scoring tools
The fourth option is a tool outside Salesforce that calculates a score and writes it to a field. These fall into three groups:
- Scoring apps on AppExchange, which add rule builders or predictive models on top of your Salesforce data.
- Predictive and intent platforms, which score accounts and people using data from outside your CRM. 6sense, for example, publishes an intent score from 1 to 100 alongside buying stages.
- Data warehouse models, where your own analysts build a model and sync the result back.
They suit teams whose best signals live outside Salesforce, such as product usage or third-party intent. For any of them, check with the vendor on three things: which Salesforce objects it writes to, whether the reasons come back with the number, and who can change the weights.
A worked example: two leads, scored two ways
Demo data. A B2B software company sells to operations teams at companies with 200 to 1,000 employees. It runs the rules-based model from Option 1, with a fit score out of 100 and an engagement score out of 100. Salesforce formula fields have no fixed scale. The 100-point caps are only this example’s choice, counted in Salesforce points, and your own model can add up to any total you like. Two leads arrive on the same morning.
| Lead A | Lead B | |
|---|---|---|
| Who | Operations Director, 350-person manufacturer, work email | Student, personal email, no company |
| Company size (max 30) | +30 | 0 |
| Industry (max 25) | +25 | 0 |
| Title (max 25) | +25 | 0 |
| Region (max 20) | 0 (field is empty) | 0 |
| Personal email | 0 | -15 |
| Fit score | 80 | -15 |
| Demo request (+40) | 0 | +40 |
| Pricing page views (+10 each, limit 2) | +10 | +20 |
| Webinar attended (+15) | 0 | +15 |
| Engagement score | 10 | 75 |
Added into one number, Lead A scores 90 and Lead B scores 60. With a routing threshold of 50, both go to a rep, and nothing on the record says that Lead B is an active poor fit while Lead A is the right buyer who has barely looked. Kept apart, the plays are obvious: Lead A goes to a rep for outreach and into a nurture track, and Lead B gets an automated reply. The empty region field on Lead A earned nothing, which is why step 2 of the build matters.
Account Engagement makes the same split in its own terms: Lead B would carry a high score and a low grade, and Lead A the reverse.
Common mistakes with Salesforce lead scoring
| Mistake | What happens | Fix |
|---|---|---|
| One blended number | Active poor fits clear the same threshold as good fits | Keep a fit field and an engagement field |
| Turning on Einstein below the minimums | You get the global model without knowing it | Check 1,000 leads and 120 conversions in 200 days first |
| Inconsistent lead conversion | The model is trained on habits, not outcomes | Agree when a lead gets converted, then stick to it |
| Using the Account Engagement score as a fit signal | Busy visitors look like buyers | Read it with the grade |
| No decay on activity points | Last year’s webinar still counts | Reduce or reset points after inactivity |
| A score no one can explain | Reps ignore it | Put the inputs or top factors on the lead page |
| Never checking the result | The score drifts from reality | Each quarter, compare conversion by score band |
Data quality sits under most of these. In Validity’s 2025 survey, 76% of CRM users said less than half of their data is accurate and complete. Our piece on where teams get ICP scoring wrong covers the modeling errors in more depth.
How to choose
- Under the Einstein minimums, or you want to own the logic: build the rules-based score.
- Plenty of conversion history and Sales Cloud Einstein in your edition: turn on Einstein Lead Scoring, and read the top factors for a month before you route on it.
- Running Account Engagement: use the score and grade together, and tune the default rules to your own definition of a qualified lead.
- Signals outside Salesforce: look at a third-party tool, and ask to see the reasons on a real record.
Where BoostIdeal fits
This guide is for teams on Salesforce, and BoostIdeal does not connect to Salesforce. It works with HubSpot, or a CSV you upload. You build the model and set the weights, and each score, shown out of 10, comes with its reasons. In HubSpot, the score and reasons write back to HubSpot properties. It is not the right fit if your data is not in HubSpot and you do not want to work from CSV uploads, or if you want a vendor to hand over a number you never question.
The principles carry over to either CRM. Keep fit and engagement apart, show the reasons on the record, and make sure a person owns the weights. Our guide to ICP scoring covers the fit side in depth.
Frequently asked questions
Does Salesforce have lead scoring?
Yes, in several forms. Sales Cloud offers Einstein Lead Scoring with Sales Cloud Einstein, Account Engagement offers a prospect score and grade, and any team can build a rules-based score with formula fields and Flow. There is no single default lead score that every edition includes.
How does Einstein Lead Scoring work?
Einstein analyzes your past leads, including custom fields, to find what converted leads have in common, then scores current leads by how closely they match. The score appears in the Einstein Score component on the lead page along with the fields that influenced it most. Salesforce says Einstein reanalyzes lead data every 10 days and refreshes the scores.
How many leads does Einstein need?
For a model built on your own data, Salesforce requires at least 1,000 leads created in the last 200 days, with at least 120 of them converted. The requirement applies to each lead segment. Below that, and only when you score all leads together, Einstein uses a global model based on anonymous data from many Salesforce customers.
What is the difference between a score and a grade in Salesforce?
In Account Engagement, a score measures activity, such as page views and form submissions, and rises or falls as a prospect interacts with your marketing. A grade measures fit against a profile of your ideal customer, runs from F to A+, and starts at D. Read together, they show who is both interested and worth pursuing.
Which Salesforce editions include Einstein Lead Scoring?
Salesforce’s help pages list it with Sales Cloud Einstein, which is included in Performance and Unlimited Editions and available for an extra cost in Enterprise Edition. Packaging changes, so confirm with your account team.
Can I do lead scoring in Salesforce without Einstein?
Yes. A formula field with a Number return type can score fit from lead fields, and a record-triggered flow can add points to a number field when activity is recorded. You own the rules, so you also own the upkeep.
What is Einstein Behavior Scoring?
It is the AI score in Account Engagement that rates prospects from 0 through 100 based on engagement patterns. Salesforce lists it for Account Engagement Advanced and Premium Editions, and says tailored insights need six months of engagement data and at least 20 prospects linked to opportunities.
The takeaway
Salesforce gives you more than one way to score a lead, and they are not interchangeable. Rules you build are transparent and yours to maintain. Einstein Lead Scoring saves the modeling work if you have the conversion history. Account Engagement separates activity from fit with a score and a grade. Start with the option your data can support, keep fit and engagement as two numbers, and put the reasons where reps can read them.