Lead Scoring Accuracy Rebuild

Results

Objective

The existing company scoring model in HubSpot was not calibrated against what CSMs actually considered target accounts. Leads were being scored using generic criteria, with no input from the people who knew which companies mattered. The result was inaccurate scores and missed opportunities.

The goal was a model built around real business priorities, validated against CSM data, and designed to be iterated on as the business evolved.

How it was built

1. Requirements gathering with Sales Managers, CSMs, and RevOps to identify the key scoring attributes: Industry, Revenue, and Headcount
2. Built a Google Sheets matrix so stakeholders could experiment with different attribute weights before anything was built in HubSpot
3. Consolidated feedback to finalise scoring tiers and attribute weights
4. Built the new scoring model using custom calculation properties in HubSpot
5. Created a dedicated HubSpot view showing old scores, new scores, and CSM target account scores side-by-side for direct comparison
6. Iterated based on feedback: excluded certain industries, refined using the Product field for better industry classification, adjusted revenue-based scoring criteria
7. Validated the final model by extracting data to Google Sheets and comparing old vs. new performance using Mean Absolute Error (MAE)

Key outcomes

• 30% improvement in scoring accuracy, measured against CSM target account data using MAE
• Model built collaboratively with stakeholders, grounded in real business priorities rather than generic criteria
• Scoring logic documented and explainable, making future iteration straightforward