Not all leads are equal. Some will buy next week; others will never convert. Odoo's predictive lead scoring uses machine learning to analyse your historical sales data and assign a probability score to every lead in your pipeline. The result: your sales team spends time on prospects most likely to close, instead of working through the list top to bottom.
This guide covers how to enable, configure, and get real value from predictive lead scoring in Odoo CRM.
Step 1: Check prerequisites
- You need Odoo Enterprise — predictive lead scoring is not available in the Community edition.
- The CRM module must be installed and actively used. The AI model learns from your historical data, so you need a reasonable volume of won and lost opportunities.
- As a rough guide, you need at least 100–200 closed opportunities (a mix of won and lost) for the model to produce meaningful scores. The more data, the better the predictions.
- Ensure your CRM data is clean: opportunities should have accurate won/lost statuses, realistic expected revenue figures, and populated contact fields (country, industry, source, etc.).
If you've just started using Odoo CRM, you won't have enough historical data for predictive scoring yet. Use manual scoring rules (covered later in this guide) as a stopgap until the AI model has enough data to learn from.
Step 2: Enable predictive lead scoring
- Navigate to CRM → Configuration → Settings.
- Under the CRM section, find Predictive Lead Scoring and enable it.
- Click Save.
- Odoo will begin analysing your historical opportunity data. Depending on the volume of data, this may take a few minutes to several hours.
- Once complete, every lead and opportunity in your pipeline will display a probability score (0–100%) representing the likelihood of winning that deal.
Step 3: Understand how the scoring model works
Odoo's predictive model examines patterns in your closed opportunities to identify which characteristics correlate with winning deals:
- Lead source — Do referrals convert better than cold outreach? Do website enquiries close faster than trade show leads?
- Country and language — Are leads from certain regions more likely to convert?
- Tags and categories — Do leads tagged with specific interests or industries win more often?
- Pipeline stage velocity — How quickly does a lead move through stages? Deals that stall in early stages tend to lose.
- Email and phone data — Does the lead have a corporate email (more likely to be a genuine business enquiry) vs. a free email provider?
- Days since creation — Older leads that haven't progressed have a lower probability.
- The model retrains automatically as new opportunities are won or lost, so scores improve over time.
Step 4: Configure scoring variables
- Go to CRM → Configuration → Settings → Predictive Lead Scoring.
- Select the fields that the model should consider. By default, Odoo uses source, country, language, and tags. You can add or remove fields based on what's relevant to your business.
- Add custom fields if you track industry-specific data — for example, company size, budget range, or product interest.
- Click Update Probabilities to retrain the model with your chosen fields.
- Review the scores on a sample of recent opportunities. Do the high-probability leads look like genuine hot prospects? Do the low-probability ones match your intuition about which deals are unlikely to close?
Adding too many scoring variables can cause overfitting — the model finds spurious patterns in small datasets. Start with 4–6 key variables and add more only if you have a large volume of historical data (500+ closed opportunities).
Step 5: Set up manual scoring rules as a complement
Predictive scoring works best alongside manual rules that capture business logic the AI can't learn from data alone:
- Go to CRM → Configuration → Scoring Rules (if available in your Odoo version) or use Automated Actions to implement manual scoring.
- Create rules that add or subtract points based on specific criteria. Examples:
- +20 points if the lead's expected revenue exceeds $50,000.
- +15 points if the lead came from a referral or partner.
- +10 points if the company has more than 50 employees.
- -10 points if the lead has no phone number (harder to reach).
- -20 points if the lead has been in the "New" stage for more than 30 days without activity.
- These manual rules combine with the AI's predictive score to give you a comprehensive picture.
Step 6: Use scores to prioritise your pipeline
- In the Pipeline view, sort or filter opportunities by probability score. Focus your daily efforts on the highest-scoring leads.
- Create a custom filter: "Probability > 60%" to see only the hot leads at a glance.
- Set up pipeline views per priority tier: Hot (70–100%), Warm (40–69%), Cold (0–39%).
- Use the Activities view to see which high-probability leads have overdue follow-ups — these are your most urgent actions.
- In team meetings, review the top 10 highest-scoring leads and ensure each has a clear next step assigned.
Step 7: Automate actions based on lead score
- Go to Settings → Technical → Automated Actions and create rules triggered by score thresholds.
- High-score automation: When probability exceeds 70%, automatically assign the lead to a senior salesperson, send an internal notification, and schedule a phone call activity within 24 hours.
- Low-score automation: When probability drops below 20%, move the lead to a nurturing email sequence instead of active sales follow-up.
- Score change alerts: Notify the assigned salesperson when a lead's score increases significantly (e.g., jumps from 30% to 60%) — something changed that makes this lead worth pursuing.
- These automations ensure your team's response matches the lead's potential without manual triage.
Step 8: Integrate scoring with lead assignment
- Go to CRM → Configuration → Sales Teams.
- Configure lead assignment rules that factor in lead score. High-scoring leads go to your best closers; lower-scoring leads go to the team handling nurture and qualification.
- If you use round-robin assignment, consider weighting it so top performers receive a higher proportion of high-probability leads.
- For businesses with territory-based assignment, use lead score as a secondary criterion — within a territory, higher-scoring leads get assigned first.
Step 9: Report on scoring effectiveness
- Go to CRM → Reporting → Pipeline Analysis.
- Group opportunities by probability range and compare actual win rates. If leads scored 70–100% are winning at 60–80%, the model is working well. If there's no correlation between score and win rate, revisit your scoring variables and data quality.
- Track conversion rates by score tier over time. The model should improve as more data accumulates.
- Measure sales cycle length by score tier — high-scoring leads should close faster on average.
- Report on revenue per score tier to demonstrate the dollar value of focusing on high-probability leads.
Step 10: Maintain and improve the model
- Mark opportunities as won or lost promptly. The model only learns from closed deals — if your team leaves stale opportunities sitting in the pipeline, the model has less data to work with.
- Record loss reasons. When marking an opportunity as lost, select a reason (price, competitor, timing, etc.). This enriches the dataset and helps the model identify which loss patterns to watch for.
- Review scoring accuracy quarterly. Compare predicted probabilities against actual outcomes and adjust scoring variables if needed.
- Clean your CRM data regularly. Duplicate contacts, missing fields, and incorrect opportunity values all degrade model accuracy.
- Retrain manually if needed. After major changes (new product lines, entering a new market, or restructuring your sales team), go to Settings and click Update Probabilities to force a model refresh.
Predictive lead scoring is a tool, not a replacement for sales judgement. Use it to prioritise effort, but don't ignore a low-scoring lead if your salesperson has genuine rapport or insider knowledge about the account.
AI-powered lead scoring in Odoo CRM takes the guesswork out of pipeline management. When configured properly and fed clean data, it helps your sales team focus on the deals most likely to close — and that means more revenue with less wasted effort. If you need help setting up or fine-tuning lead scoring in Odoo, get in touch.