Lead Scoring Models

The problem lead scoring solves

A B2B marketing funnel generates leads of wildly uneven quality — a VP at a target-account company who requested a demo is not the same as a student who downloaded a free ebook out of curiosity. Sending both directly to a sales rep wastes sales time on the second lead and, at scale, trains reps to distrust marketing-sourced leads entirely. Lead scoring is the system that assigns each lead a numeric value representing sales-readiness, so that only leads crossing a defined threshold get routed to sales, while everyone else continues in nurture.

Explicit vs. implicit scoring

Lead scoring models combine two categories of signal:

- Explicit (demographic/firmographic) scoring — attributes the lead tells you directly or that can be inferred from their profile: job title, company size, industry, geography. This answers "is this lead who we want to sell to at all?" (fit), independent of their behavior.
- Implicit (behavioral) scoring — actions the lead takes: pages visited, content downloaded, emails opened/clicked, webinar attendance, pricing page visits, repeat site visits. This answers "how interested and how close to a buying decision is this lead?" (intent).

A lead can score high on fit and low on intent (a perfect-profile prospect who's barely engaged — needs nurture, not a sales call) or high on intent and low on fit (a highly engaged student or a competitor doing research — needs to be filtered out, not routed to sales). Combining both axes into a single score prevents either failure mode from being missed.

Explicit vs. implicit scoring signals

Dimension	Explicit (fit)	Implicit (intent)
Source	Form fields, enrichment data (Clearbit, ZoomInfo)	Behavioral tracking, email engagement, CRM activity
Examples	Job title = 'VP Marketing', company size 200+	Visited pricing page 3x, attended demo webinar
Answers	Should we sell to this account at all?	Are they close to a buying decision?
Decays over time?	No — relatively static	Yes — recent activity weighted higher

Building the point model

Most practical scoring models are additive point systems: each qualifying attribute or action adds (or subtracts) points, and the running total determines lifecycle stage. Negative scoring matters as much as positive — a personal Gmail address, a student job title, or a competitor's company domain should actively subtract points, since without negative signals disqualifying leads can accumulate enough positive behavioral points to falsely cross the sales-ready threshold.

lead_scoring_rules.txt

text

EXPLICIT (fit)
  Job title contains "Director"/"VP"/"Head of"     +20
  Job title contains "Manager"                      +10
  Job title contains "Student"/"Intern"             -30
  Company size 200+ employees                       +15
  Company size 1-10 employees                        +5
  Email domain is free provider (gmail, yahoo)      -15

IMPLICIT (intent)
  Visited /pricing                                  +10
  Visited /pricing 3+ times in 7 days                +15
  Requested demo                                     +25
  Downloaded ebook / gated whitepaper                 +5
  Opened 3+ marketing emails, no clicks               +2
  No site visit in 45 days                          -10 (decay)

THRESHOLDS
  0-39   = MQL-pending (continue nurture)
  40-69  = MQL (Marketing Qualified Lead → nurture + sales visibility)
  70+    = SQL (Sales Qualified Lead → routed to sales queue)

MQL and SQL are handoff contracts, not universal definitions

There's no industry-standard point value that makes a lead an MQL versus an SQL — the thresholds only mean something as a negotiated agreement between marketing and sales about what 'ready to contact' looks like for this specific business. A threshold copied from a blog post or another company's playbook, without sales validating it against leads that actually closed, produces a score that looks rigorous but predicts nothing.

MQL to SQL: the handoff isn't automatic

Crossing the MQL threshold typically triggers marketing-side actions — sales-team notification, addition to a high-touch nurture track — but the MQL-to-SQL transition usually still requires either sales acceptance (a rep reviews and confirms fit before formally taking ownership) or a secondary qualification step (BANT/MEDDIC-style discovery call). Treating MQL crossing as equivalent to guaranteed SQL status, and auto-routing every MQL straight into a sales rep's queue without any review layer, is the most common cause of sales teams losing trust in marketing-sourced leads.

Connecting scoring to lifecycle-stage automation

Lead score should function as an input variable to marketing automation workflows, not a static report checked manually. In HubSpot, Marketo, or similar platforms, score threshold crossings typically trigger: lifecycle stage field updates (Lead → MQL → SQL), enrollment/removal from nurture email flows, internal Slack or CRM task notifications to the owning rep, and changes to lead routing rules (which rep or team the lead is assigned to, often based on scored value plus territory or account-tier rules).

Because implicit score should decay over time (a lead who was highly engaged three months ago but has gone silent shouldn't remain permanently marked sales-ready), most mature implementations run a recurring recalculation job — nightly or on every new engagement event — rather than scoring leads once at initial capture.

Static one-time scoring stales fast

A lead scored once at form submission and never re-evaluated becomes inaccurate almost immediately — the score reflects a moment, not an ongoing relationship. Without decay and recalculation, marketing keeps routing leads to sales based on interest that cooled months ago, and genuinely re-engaged older leads never get flagged at all.

What's next

Lead scoring depends entirely on having clean, unified behavioral and firmographic data attached to each contact record — which is the job of the underlying CRM data model.

Next: CRM Data Model →