Building the Composite Score

A composite intent score aggregates multiple signal types into a single number that tells your demand generation team: "This account is worth pursuing right now." The score must balance signal diversity (more signal types equals higher confidence), signal recency (recent signals are more predictive than stale ones), and signal strength (a pricing page visit is stronger than a blog view). The model should be simple enough to explain and validate, but nuanced enough to differentiate truly in-market accounts from noisy data.

1. Define Signal Categories Group all available signals into categories: first-party behavior (website visits, content engagement), third-party research (topic consumption, surge data), and market signals (hiring, technology changes, trigger events). Each category contributes independently to the composite score. An account with signals in all three categories is more likely to convert than an account with strong signals in only one.
2. Assign Base Weights Weight each signal by its historical correlation with pipeline creation. If pricing page visits predict meetings at twice the rate of blog views, weight them accordingly. Start with estimated weights based on experience, then calibrate against actual conversion data after 90 days.
3. Apply Decay Functions Intent signals lose predictive value over time. A website visit from yesterday is highly predictive. A website visit from 60 days ago is noise. Apply a decay function that reduces signal weight as time passes — full weight for signals within seven days, 50% weight for 8-30 days, 25% weight for 31-60 days, zero weight after 60 days.