Rules get set once and sit for quarters. Small scoring misses stack up, and by the time Q3 buying cycles hit, you're leaking deals you should win.
Treating Every Pageview Like a Buying Signal
A home page visit isn't pricing. A careers page isn't a buying signal at all. One SaaS team gave points for any 3-page session — 60% of MQLs turned out to be students, competitors, and lost visitors. Once they focused scoring on product, pricing, and integration pages, MQL volume dropped ~35% but meetings per MQL doubled.
Group pages by intent tier and only score the upper tiers:
- Learn: blog, generic resources, product overview
- Evaluate: features, comparison, integration detail
- Buy: pricing, demo, checkout, quote
Ignoring Identity
Treating every anonymous session as net-new gives "new lead" points to your own customers on a different device. In one audit, 20% of new MQLs were existing customers or open-opp contacts. Even a basic identity spine across devices, emails, and accounts can cut duplicate "net new" leads by 15–30%.
Overweighting Form Fills, Underweighting Behavior
A newsletter signup is not a "talk to sales" request. Calibrate to intent strength and combine with behavior — someone who hits pricing twice in three days should outrank a one-time demo from a generic email. Your score gap should mirror real conversion ratios:
- Low: ebook, webinar, newsletter
- Medium: trial, sample, tool signup
- High: talk to sales, pricing, quote, demo
Static Scores Instead of Time-Bound Intent
Scores that only go up create stale funnels. Add recency bands (1, 7, 30 days), weight recent actions more, and subtract points for 30–90 days of inactivity. One mid-market SaaS team reclassified 40% of "hot" leads as dormant in a month — meetings per outreach went from ~4% to 8%.
Disconnecting Scoring From Routing
Territory-only or pure round-robin routing undoes good scoring. Tie scoring, segments, and workflows together:
- Different SLAs by score and role
- Target accounts and senior titles to experienced reps
- Separate flows for customers and partners vs. net-new
- Pass clear behavior fields into CRM so sales sees what happened
If high-score visitors wait 15–30+ minutes for human follow-up during business hours, routing is fighting your scoring.
Make It a Testable System
Each quarter, compare MQL→opportunity by source and score band, study your last 50 closed-won deals, and test one change per month — new decay rules, updated page weights. Track MQL volume and MQL→opportunity over the next 30–60 days.
Next Step
Map your current model this week and circle three rules that feel off. Then see how our website visitor lead scoring tightens it, or book a demo.