Einstein Lead Scoring: AI Predictive Scoring

Effective Lead Management is essential for a healthy sales pipeline. Prioritisation is key. The end goal is to work with the best leads, so they become happy customers (and businesses gain a return on investment). 

This post overviews Einstein Lead Scoring within Salesforce Lightning Experience. 

Screenshot of Einstein Lead Score
Screenshot of Einstein Lead Scoring

Table of Contents


If you are familiar with Leads, Lead Scoring (principle of) and Einstein AI, please feel free to skip to Einstein Lead Scoring: Overview.

What is a Lead?

A Lead is someone who expresses an interest in your products or services. For example, it could be someone wanting to buy a laptop. It could also be a business wanting to buy 1,000 laptops. Whatever your market, you need to capture this interest. 

Why Lead Scoring?

Imagine you work in Sales. Each day, you receive queries from (hopefully many good quality!) prospective clients. You have to engage and qualify your existing Leads. There is never enough time in the day. If only there was a way to prioritise the leads… 

This is where Lead Scoring comes in. Lead Scoring provides a rating for each Lead, allowing you to focus your efforts on the ‘best’ ones.

Why Einstein (AI)?

You may be wondering what this has to do with Einstein. Salesforce Einstein is a collection of Artificial Intelligence (AI) solutions, built into Salesforce. 

Einstein Lead Scoring is one offering, available in ‘Sales Cloud Einstein’Salesforce analyses (machine learning) your Lead data to identify characteristics of ‘successful’ (converted) Leads. Salesforce then gives Leads a score of 0-100, helping with prioritisation.  This focuses efforts on prospects who are most likely to become customers. 

Einstein Lead Scoring vs 'Criteria-based' Lead Scoring

Always consider the problem being solved. Using criteria/rule-based scoring (e.g. Formula fields/automation) could be an alternative. With this, you define the rules and ‘calculate’ a score. Before you do, ask yourself this: 

  • Do you objectively know what factors matter most when qualifying/converting Leads?
  • Do you have the knowledge/resources to monitor changes in customer traits over time? 
  • Is your sales process simple/easy to predict?

If you feel the answer is ‘Yes’ to the above, you may be able calculate scores. However, if ‘No’, AI may help identify underlying trends. The predictive model re-evaluates every 10 days and refreshes scores, providing up to date insights. 

NOTE: Einstein Lead Scoring does not update scores ‘in real-time’. When applicable, scores are updated every few hours. Click here for more info.

Einstein Lead Scoring: Overview


There are some prerequisites to consider as of Winter ’21. These are outlined below. 


You need the following:

Data Requirements

For Machine Learning to work, there needs to be sufficient data to create a model. This model allows the system to generate a ‘Lead Score’. There are two modes available outlined below. 

  • Own predictive model
    • For a custom predictive model, the following data requirements must be met:
      • A minimum of 1,000 leads created in the past 6 months (180 days)
      • At least 120 of these must have been converted to an Account and Contact OR Account, Contact and Opportunity

Tip: It is not just ‘data quantity’ that matters. Quality does too. Poor or insufficient data will skew predictions.  

  • Global Model
    • If your Salesforce org does not meet these conditions, a ‘global’ model is applied. This uses anonymised data from many Salesforce Customers to produce a prediction. 

For a more detailed outline of requirements and considerations, please click here

Setup - Helpful Resources

Salesforce provides ‘Assisted Setup’ for Einstein Lead Scoring. For instructions, please refer to these resources:

Here as some useful tips/pointers:

  1.  Do you want to include all of your Leads?
    • For example, you may have a high-volume of low quality data or very different sales leads. Could these be distinguished by a field or record type? Consider whether including these would skew your model’s output 
  2.  Are there fields you wish to exclude from the model?
    • Salesforce recommends including all fields. If you are confident a field does not impact Lead conversion chances, consider excluding   
  3. Einstein Lead Scores do not refresh instantly
  4.  Plan how to use the scores
    •  How will you use this to help users? See the next section for ideas

Outputs & Use Cases

Users are able to see the ‘Lead Scores’ in various places. For example, the scores can be easily shown within a List View (below) or in Reports:

Screenshot of Einstein Lead Scores on a List View
Screenshot of Einstein Lead Scores on a List View

You can also add the ‘Lead Score’ field to the page, as well as the ‘Einstein Scoring’ component. The component allows you to see factors which positively and negatively influenced the score, where eligible:

Screenshot of Einstein Scoring Lead Component
Screenshot of Einstein Scoring Lead Component

Lead Scores can also be used to drive automation. For example, sending emails/notifying user, assigning Leads, etc. In short, Lead Scoring has the potential to be a powerful addition to Lead Management processes. 


Einstein Lead Scoring is an add-on product. Its machine learning can help identify traits in previously successfully converted leads. This insight can help sales users prioritise higher quality Leads. 

The product can be easily tied into List Views, Pages, Reports and automation. Consider how it may suit your use case and any factors which may impede adoption (e.g. see Prerequisites and Setup – Helpful Resources). 

Bonus Penguin Fact

Penguins have been around a very long time. In 2010, a fossil of a giant penguin Inkayacu paracasensis (otherwise known as the Water King) was found, which dated back to 36 million years ago (source: BBC, 2010). Fossils have shown this penguin to be 1.5 meters tall! 

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.