Predictive Lead Scoring
What is Predictive Lead Scoring?
Predictive Lead Scoring replaces hand-tuned scoring rules with a model that learns from data. Inputs include firmographics, technographics, engagement signals, and contextual factors; the model outputs a probability or score. The advantage is that models surface patterns humans wouldn't catch — 'company size between 200-400 employees, using Salesforce, with VP-level engagement in the last 14 days' might be the strongest predictor, and a model finds it. The disadvantages: models need historical data to train (a problem for early-stage companies), and they drift as the business evolves.
Why it matters
- Surfaces conversion patterns that hand-tuned rules miss.
- Improves with more data — models get more accurate as the company grows.
- Requires monitoring — models drift, and a stale model is worse than a clear rule.
Use cases
- Mid-funnel prioritization. ML model scores leads on close-likelihood; high-score leads route to senior AEs.
- Re-engagement timing. A model predicts which dormant leads are likeliest to re-engage.
- Sales-marketing alignment. Shared model reduces debate over which leads are 'good'.
How turgo helps
turgo combines rule-based ICP scoring with predictive models trained on the customer's own win history — getting the interpretability of rules and the lift of machine learning.
See turgo in action →
