Why Most Lead Scoring Fails
Most lead scoring models are built by marketing teams who've never carried a quota. They assign points for page views and email opens, then wonder why sales ignores the 'hot leads.'
The fundamental problem is confusing activity with intent. Someone who reads 10 blog posts might be a researcher or competitor. Someone who visits your pricing page and downloads a case study is probably shopping.
The Intent-Based Framework
At Selworthy, we build lead scoring models around three dimensions: fit, intent, and engagement.
Fit Score (Demographic)
Does this person match your ICP? Score based on company size, industry, role, and technology stack. No amount of engagement should override a terrible fit score.
Intent Score (Behavioral)
We weight actions by purchase intent: pricing page visits (high), case study downloads (high), blog reads (low). Recency matters more than frequency.
Engagement Score (Relationship)
Has this person interacted with your sales team? Attended a webinar? These relationship signals indicate active evaluation.
Implementation in HubSpot
Step 1: Define Your Thresholds
Work with sales to define what score qualifies as an MQL, SQL, and 'not ready.' Base it on historical conversion data.
Step 2: Build the Model
Create separate HubSpot score properties for fit and intent. Use calculated properties to combine them.
Step 3: Test and Iterate
Run the model in shadow mode for 30 days. Adjust weights until the model's top decile converts at least 3x the average.
Common Pitfalls
Score inflation is the biggest killer. Implement score decay — points should expire after 30-60 days of inactivity. The best models are simple enough that a sales rep can explain why a lead scored high.
