How to Set Up Lead Scoring in Odoo 19: Predictive & Manual

Lead Scoring in Odoo 19

Odoo 19 offers two approaches to lead scoring: predictive (machine learning) and manual (rule-based). Both help sales teams focus on the leads most likely to convert.

Predictive Lead Scoring (ML-Based)

How It Works

Odoo analyzes your historical won/lost opportunities and builds a model that predicts conversion probability for new leads. The score (0-100%) appears on each lead/opportunity.

Setup

1. Go to CRM → Configuration → Settings
2. Enable Predictive Lead Scoring
3. Select scoring variables:
   • Country — geographic patterns
   • State — regional patterns
   • Email Quality — domain analysis
   • Phone Quality — phone number presence
   • Language — language preference
   • Team — sales team performance
4. Click Save

Requirements

• Minimum 50 won and 50 lost opportunities for the model to train
• More historical data = better predictions
• Model retrains automatically as you close more deals

Using Predictive Scores

In the pipeline kanban view, leads show their probability score. Sort by probability to focus on high-scoring leads first. The score updates automatically as lead data changes.

Manual Lead Scoring (Rule-Based)

For teams without enough historical data, or for additional scoring criteria:

Using Tags for Scoring

# Create tags with score values:
# Hot Lead (score: 90)
# Warm Lead (score: 60)
# Cold Lead (score: 30)

# Apply tags manually or via automation rules

Using Automated Actions

1. Go to Settings → Technical → Automated Actions
2. Create rules that update the probability based on criteria:
   • If country = USA → set probability = 70%
   • If source = Website Form → set probability = 60%
   • If email domain = gmail.com → set probability = 40%

Score-Based Lead Assignment

Combine scoring with assignment rules:

1. High-score leads (>70%) → assign to senior salespeople
2. Medium-score leads (40-70%) → assign to inside sales team
3. Low-score leads (<40%) → assign to nurture campaign

Best Practices

• Wait for data — Predictive scoring needs at least 100 closed opportunities (50 won + 50 lost)
• Review regularly — Check if high-scoring leads actually convert. Adjust variables if not.
• Don't ignore low scores — Use them for nurture campaigns, not deletion
• Combine approaches — Use predictive + manual rules for best results