Einstein Lead Scoring Setup: Step-by-Step Implementation Guide for Sales Teams

You've been burned before. A hot lead comes in from a high-intent source - the company's the right size, the job title matches your ICP perfectly. But it sits in a queue for four hours because your routing can't tell the difference between that lead and a student who just downloaded a whitepaper.

‍

That's the problem Einstein Lead Scoring solves. Instead of assigning every lead the same priority, Salesforce Einstein scores each one by its actual likelihood to convert - based on your own historical data. And when you connect those scores to your routing logic, your best leads get to your best reps, fast.

‍

This guide walks through exactly how to set up Salesforce Einstein Lead Scoring, what to do with the scores once you have them, and how to use them as the foundation for a lead prioritisation system that actually works.

‍

What Is Einstein Lead Scoring in Salesforce?

Einstein Lead Scoring is a predictive AI feature built into Salesforce Sales Cloud that analyses patterns in your historical lead and opportunity data to assign each new lead a score between 0 and 100. A higher score means a higher predicted likelihood to convert.

‍

Unlike rules-based scoring - where you manually assign points for actions like form fills, email opens, or job title matches - Einstein Lead Scoring is statistical. It looks at which past leads became won opportunities and identifies the shared characteristics that predicted that outcome. Then it applies those patterns to new leads in real time.

‍

Critically, Einstein is learning from your data, not generic benchmarks. A lead scoring model that works for a fintech selling to CFOs looks very different from one that works for a SaaS company targeting IT managers. Einstein builds that model for you automatically.

‍

If you don't yet meet the data threshold (more on that below), Einstein will fall back to a global model built from anonymised Salesforce customer data. You'll still get scores - they just won't be as tailored as an org-specific model. Once your data catches up, Einstein switches over automatically.

Einstein Lead Scoring vs Rules-Based Lead Scoring

Both have a place in a mature Salesforce lead scoring setup.

‍

Rules-based scoring (e.g. Pardot/Account Engagement scoring) is good for tracking explicit signals - job title, page visits, email engagement, form submissions. You define the logic; it runs it.

‍

Einstein Lead Scoring is good for predicting conversion likelihood from patterns across dozens of implicit signals, including combinations of fields that human logic wouldn't catch. The model updates automatically as your win rates evolve.

‍

The most effective lead scoring strategy in Salesforce uses both: Einstein for the headline score, and rules-based signals to segment within bands or add context for routing.

‍

Prerequisites: What You Need Before You Set Up Einstein Lead Scoring

Before you start, check you have the following in place.

‍

Salesforce Edition. Einstein Lead Scoring is included in Sales Cloud Unlimited and Einstein 1 Sales / Unlimited+editions. It's available as a paid Sales Cloud Einstein add-on for Enterprise and Performance editions (typically $50/user/month). It is not available on Professional or Essentials.

‍

Data requirements. You need a minimum of 1,000 lead records created in the last 200 days, and at least 120 of those leads converted to an account and contact in the same window, for Einstein to build an org-specific model. Below that threshold, Einstein will use the global fallback model instead of an org-specific one.

‍

Permissions. You need System Administrator access to enable Einstein Lead Scoring in Setup. Individual users need the 'View Einstein Lead Scores' permission to see scores on records.

‍

Platform encryption. Einstein Lead Scoring is not currently compatible with Salesforce Shield platform encryption. Worth flagging if you're in financial services or another regulated vertical where Shield is in play.

‍

Lead and Opportunity data quality. Einstein's model is only as good as your data. If your leads have inconsistent job titles, missing company fields, or conversion events that don't reflect real pipeline, your scores will reflect that noise. A basic data hygiene audit before setup is time well spent.

‍

How to Set Up Einstein Lead Scoring: Step-by-Step

Step 1: Enable Einstein Lead Scoring

1. Go to Setup > Einstein Sales > Einstein Lead Scoring.
2. Click 'Get Started' (or 'Enable' depending on your Salesforce version).
3. Review the data sharing prompt and confirm.
4. Click Enable.

‍

Once enabled, Einstein begins analysing your data immediately. Initial scoring typically takes up to 48 hours depending on your data volume.

Step 2: Configure Your Scoring Model

• Define the conversion event. By default, Einstein looks at 'IsConverted = true' on Lead records. If your team uses a custom conversion milestone (e.g. opportunity creation at conversion), configure this here.
• Exclude fields if needed. Go to Setup > Einstein Lead Scoring > Model Settings and toggle off any fields that contain irrelevant noise (internal tracking codes, legacy fields no one fills in).
• Consider segmented scoring. If your lead population is genuinely bimodal - e.g. inbound web leads and partner-referral leads that convert through very different patterns - Einstein supports separate scoring models per segment. Each segment needs to independently meet the minimum data threshold, but it prevents one model trying to average across two incompatible conversion patterns.

‍

Step 3: Add the Einstein Lead Score Field to Page Layouts

1. Go to Setup > Object Manager > Lead > Page Layouts.
2. Edit the layout used by your sales team.
3. Search for 'Einstein Lead Score' in the fields panel.
4. Drag it onto the layout - ideally at the top of the record, near Status and Owner.
5. Also add the 'Einstein Lead Score Insights' component (showing Top Positives and Top Negatives) so reps can see why a lead is scored the way it is.
6. Save.

Step 4: Add Einstein Scores to List Views and Reports

List Views: Go to Leads > List Views > Edit Columns and add 'Einstein Lead Score'. Create a dedicated view sorted by score descending - 'High Priority Leads' is the obvious name.

‍

Reports: Add Einstein Lead Score as a column to your lead pipeline reports. Bucket scores into bands (0–25, 26–50, 51–75, 76–100) using formula fields for cleaner reporting.

‍

Dashboards: Build a dashboard component showing lead distribution by score band. This is the clearest signal to RevOps that the model is working - or not.

Step 5: Create Score-Band Classification Using Custom Fields

Create a new custom field on Lead: Formula (Text), name it 'Lead Priority Band', then use the following logic:

‍

• Score 75–100 = P1 - High
• Score 50–74 = P2 - Medium
• Score 25–49 = P3 - Low
• Score below 25 = P4 - Nurture

‍

This classification field is what you'll reference in Distribution Engine to set routing rules and SLAs per band - keeping your routing logic clean and maintainable without needing to write range conditions.

‍

Building Your Lead Scoring Criteria: What Signals Matter

Einstein's model identifies scoring signals automatically, but understanding which signals it values most helps you make better decisions about your data, routing, and ICP definition.

‍

Example scoring signals and their impact:

‍

• Job Title = VP/Director. High authority, strong ICP match. Score direction: High Positive.
• Industry = Finance/SaaS. Core verticals with demonstrated ROI. Score direction: High Positive.
• Company Size = 200–2,000 employees. Mid-market sweet spot. Score direction: Positive.
• Visited Pricing Page. High purchase intent signal. Score direction: High Positive.
• Opened 3+ Emails. Active engagement, nurtured. Score direction: Positive.
• Job Title = Intern/Student. Out of ICP. Score direction: High Negative.
• Free Email Domain. Low conversion probability. Score direction: Negative.
• No Activity in 30 Days. Cold lead, low intent. Score direction: Negative.

‍

Einstein surfaces its top factors as 'Top Positives' and 'Top Negatives' on each lead record. After your model has been running for a few weeks, audit these factors across your highest-scoring leads to sense-check that they reflect genuine ICP signals - not noise.

Combining Einstein Scores with Intent Data

Einstein scores leads based on fit and historical patterns. But intent data - signals like visiting your pricing page, attending a webinar, or engaging with high-value content - adds real-time purchase readiness to the picture.

‍

To route leads based on intent data in Salesforce, the most practical approach is to capture intent signals as custom fields (e.g. 'Pricing Page Visit = TRUE') and include them in your routing criteria alongside the Einstein score band. In Distribution Engine, you can combine score band with intent flags as conditions on a distributor - so a P2-scored lead who visited your pricing page gets treated with P1 urgency.

‍

How to Route Leads by Lead Score in Salesforce

Getting Einstein Lead Scoring set up is step one. Connecting those scores to your lead routing logic is where you actually see the revenue impact.

Option 1: Native Salesforce Assignment Rules (Basic)

Salesforce's built-in assignment rules let you add criteria based on field values, including your Lead Priority Band formula field. You can create separate rules for P1 leads (route to senior AE queue) vs P3 leads (route to SDR queue).

‍

The limitations kick in quickly: you can't set time-based SLAs, you can't balance workloads dynamically, and you can't set caps per rep. For teams with more than 10 reps or complex territory structures, native rules become a maintenance headache fast.

Option 2: Distribution Engine - Score-Driven Routing (Recommended)

Distribution Engine by NC Squared is purpose-built for score-driven routing in Salesforce. It reads your Einstein Lead Score band field and lets you configure exactly how each tier is handled. As a starting point:

‍

• P1 - High (Score 75–100). SLA target: under 10 minutes. Routing: Senior AEs, round robin, active cap of 5.
• P2 - Medium (Score 50–74). SLA target: under 1 hour. Routing: Full rep pool, weighted by seniority.
• P3 - Low (Score 25–49). SLA target: under 4 hours. Routing: SDRs, standard round robin.
• P4 - Nurture (Score below 25). No SLA. Routing: Nurture queue or marketing automation.

‍

Tune these thresholds and SLAs to your own conversion patterns - the bands above are a typical setup, not a universal rule.

‍

To prioritise high-value leads in Salesforce routing, set up separate distributors in Distribution Engine - one per score band - each with their own rep pool, distribution method, and SLA timer.

‍

To add weighting and capping on top of round robin routing, use Distribution Engine's Load Balancing with Caps. Set higher active caps for senior reps and lower caps for new-starters, so high-score leads always land with reps who have capacity and experience.

‍

The entire setup is no-code and admin-configurable. RevOps teams can change band thresholds, swap rep pools, or adjust SLAs without touching Salesforce Flow or writing Apex.

‍

Lead Scoring Best Practices for Salesforce

‍

1. Let the model stabilise before acting on it

Einstein's first model is based on whatever historical data you have at the point of setup, and it retrains every 10 days. Give it 4–6 weeks (so, at least 4 retraining cycles) of new lead activity before using scores to gate routing or SLAs. Check the model health dashboard in Setup regularly.

2. Close the feedback loop

The model improves when reps work leads properly. Einstein rescores leads at least every 6 hours (sooner if a key field changes) and retrains the underlying model every 10 days - but only if the conversion data it's learning from is clean. Enforce lead status hygiene: require reps to update to 'Working', 'Qualified', or 'Disqualified' within SLA, and auto-reassign leads that breach it.

‍

3. Don't treat score as a filter, treat it as a queue prioritiser

A low Einstein score doesn't mean a lead is worthless - it means it's less likely to convert based on historical patterns. All leads should still get worked. Lead score determines the speed to lead and rep tier that works them, not whether they get worked at all.

4. Audit for bias in your historical data

If your historical conversion data has gaps - certain lead sources never properly followed up, or one product line has a poor close rate - Einstein will reflect those gaps. A quarterly audit of score factors vs actual outcomes will catch model drift early.

5. Use score band in reporting, not just routing

Key KPIs for RevOps teams to track for lead routing in Salesforce when using Einstein scores:

‍

• Conversion rate by score band (P1 vs P2 vs P3)
• Average time-to-first-touch by score band
• SLA attainment rate by band
• Auto-reassignment rate (signals reps are not meeting SLAs)
• Model accuracy: percentage of P1 leads that converted vs P4

‍

Common Issues and How to Fix Them

Issue: Scores are not appearing on lead records

‍

Most likely cause: The model is still training (wait up to 48 hours), or the field hasn't been added to the page layout.

‍

Fix: Check Setup > Einstein Sales > Einstein Lead Scoring for model status. If status is 'Not Enough Data', you haven't met the 1,000 leads / 120 conversions threshold in the last 200 days - you'll be scored against the global model instead.

‍

Issue: All leads are scoring high (or all low)

‍

Most likely cause: The conversion event is misconfigured, or your historical data is skewed.

‍

Fix: Review your conversion event definition in Setup. Check whether your historical data contains a representative mix of converted and unconverted leads.

‍

Issue: Reps don't trust the scores

‍

Most likely cause: Scores are visible but context is not. If reps can't see why a lead is scored a certain way, they'll ignore the number.

‍

Fix: Add the Einstein Lead Score Insights component to the lead page layout so Top Positives and Top Negatives are visible. Run a short enablement session showing reps what the top factors are.

‍

Issue: Score and actual conversion are diverging over time

‍

Most likely cause: Model drift. Your ICP or product has shifted, but the model is still learning from older patterns within the 200-day training window.

‍

Fix: Go to Setup > Einstein Lead Scoring > Model Settings, review your conversion milestone and field exclusions, and consider whether segmented scoring would better reflect how your pipeline now actually converts.

‍

How NC Squared Helps.

Einstein Lead Scoring gives you the signal. Distribution Engine by NC Squared gives you what to do with it.

‍

Without a routing layer that reads and acts on those scores, they stay in reports. Reps work their favourites. SLAs slip. Marketing's best leads go cold waiting for the right rep to pick them up.

‍

• Deeply flexible routing: Round robin, territory, skill, or capacity.
• Smart automation: Real-time assignment, SLA timers, auto-reassignment.
• Analytics: Workload, performance, SLA dashboards all in Salesforce.
• No code: Teams can own their workflows without code or devs.

‍

That's why 20,000+ users trust NC Squared to keep leads moving and revenue flowing.

‍

Einstein Lead Scoring Setup Checklist

Before you start

‍

• Confirm Sales Cloud Unlimited / Einstein 1 Sales edition, or Enterprise/Performance with the Sales Cloud Einstein add-on
• Verify 1,000+ lead records and 120+ conversions in the last 200 days
• Confirm Salesforce Shield platform encryption is not in use on Lead fields
• Run a basic lead data hygiene audit (job title, company, conversion events)

‍

Setup

‍

• Enable Einstein Lead Scoring in Setup > Einstein Sales
• Configure conversion event and field exclusions
• Decide whether to use a single model or segmented scoring
• Add Einstein Lead Score and Score Insights component to Lead page layout
• Add score column to Lead list views and reports
• Create Lead Priority Band formula field (P1–P4)

‍

Routing

‍

• Set up separate distributors in Distribution Engine per score band
• Configure SLA timers per band (P1 = 10 mins, P2 = 1 hour, etc.)
• Enable auto-reassignment on SLA breach

‍

Ongoing

‍

• Review model health and score factors monthly
• Track conversion rate by score band quarterly
• Enforce lead status hygiene to keep the feedback loop clean

‍

Takeaway.

Einstein Lead Scoring gives Salesforce admins and RevOps teams something powerful: a model that learns from your own conversion history and tells you, in real time, which leads are worth prioritising. Setup is straightforward. The hard part is what comes next - making sure those scores actually change how leads are routed, worked, and reported on.

‍

Get the scoring model live, connect it to your routing rules, close the feedback loop, and you've built a lead management system that gets smarter the more you use it.

‍