Lead scoring is the practice of ranking leads by fit and engagement signals so 📝go-to-market teams prioritize follow-up and route the best opportunities to the right rep before competitors get there.
A lead score combines two distinct questions. Fit asks whether a lead matches the 📝ideal customer profile — industry, company size, role, geography, technology stack — and is largely static. Engagement asks how actively the lead is behaving — site visits, email opens, content downloads, demo requests, pricing-page views — and changes by the hour. Traditional models assigned manual point values to each attribute and behavior, summing them into a single number that thresholds into tiers like hot, warm, and cold. The score then drives routing: who gets called first, who gets nurtured, who gets dropped.
The discipline exists to solve a scarcity problem. Sales capacity is finite, and most 📝inbound leads never convert, so spraying equal effort across every record wastes the team's most expensive resource. Scoring concentrates human attention on the leads most likely to close, which is why 📝speed-to-lead — contacting a high-scoring lead within minutes rather than hours — consistently moves win rates. Done well, scoring is the routing logic that connects 📝demand generation to the 📝pipeline; done badly, it buries good leads under noisy vanity signals and trains reps to ignore the score entirely.
Modern lead scoring has moved from static point tables toward signal-based and predictive models. Enrichment APIs supply richer fit data, intent providers surface buying signals from outside the company's own properties, and machine-learning models score leads against patterns in historical closed-won deals rather than a human's guess at point values. In technical go-to-market work, scoring is increasingly an automated pipeline — enrich, score, route — that fires in real time, making it one of the clearest places where encoding judgment into a system beats manual triage.
Lead scoring is where most teams quietly leak revenue. The static point table from 2015 is still routing leads at half the companies I see — meanwhile the signal that actually predicts a close is sitting in an enrichment feed nobody wired up.