Lifecycle

What is Lead Scoring?

B2B marketing definition, context, and 42 Agency's operator take.

Last updated: April 23, 2026

The short answer

Lead scoring is a framework for ranking leads by likelihood to buy, combining ICP fit (who they are) with engagement signals (what they do). The right model produces a defensible MQL threshold.

Definition

Lead scoring assigns numeric values to lead attributes and actions, then uses the total to decide whether and when to pass a lead to sales. A modern lead scoring model separates two axes:

• Fit score — firmographic and technographic match to ICP (who they are, statically)
• Engagement score — behavior that indicates buying intent (what they do, with decay)

Each axis can be scored independently. A lead becomes an MQL only when both axes cross their threshold. This replaces the older single-score model where a champion at a bad-fit account could accidentally trigger an MQL through pure engagement volume.

Why it matters

Bad lead scoring creates two losses: sales time wasted on poor-fit leads, and good leads missed because engagement wasn't weighted correctly. When reps start ignoring the MQL queue, the scoring model has failed even if the math still runs cleanly in HubSpot.

Good scoring is measured by conversion to opportunity, not MQL volume. If the MQL-to-opportunity rate is drifting down over time, the model needs retuning — usually decay logic, new negative signals, or fresh threshold calibration.

42's take

Most lead scoring models we inherit have three problems: too many attributes (50+ that no one can defend), no decay (engagement from 2023 still counts as hot), and no regular calibration. The fix is ruthless simplification: 10-15 fit attributes, 6-10 engagement actions with 30-90 day decay, and a quarterly recalibration based on fresh conversion data.

Single-score models are mostly dead. Two-axis models (fit × engagement) produce better pass-through rates and make the disagreement between marketing and sales resolvable with data rather than opinion.

Frequently asked questions

What signals should be in a lead scoring model?

Fit signals: industry, company size, revenue, tech stack, geography. Engagement signals: repeat website visits, high-intent page views (pricing, demo request), content downloads, webinar attendance, product usage if PLG. Exclusions: competitors, current customers, students.

How often should lead scoring be recalibrated?

Quarterly. Pull the last 90 days of MQLs, compute conversion-to-opportunity, and adjust the threshold if the rate has drifted by more than 20%. Re-weight individual attributes if certain signals stop correlating with conversion.

Should fit score and engagement score be combined into one number?

No — keep them separate. A single combined score lets a high-engagement bad-fit lead trigger an MQL. Two-axis models require both thresholds to be crossed, which matches the actual qualification intent.