Bad lead scores waste rep time on tire-kickers and skip qualified buyers. Here's the fit + intent framework that actually predicts closed-won.
Lead scoring exists to answer one question: which of today's leads should a rep call first? Most scoring models are elaborate point systems that add up to garbage — 100 points for downloading a whitepaper, 50 for attending a webinar, 25 for opening an email. The rep learns to ignore the score within a month. The scoring model that works separates two dimensions: fit (do they match our ICP?) and intent (are they in a buying window?).
Fit: static attributes of the account (industry, size, tech stack, geography) that predict whether they'll ever be a customer. Intent: behavioral signals (pricing page visits, demo requests, RFP mentions, competitor evaluation) that predict whether they're in a buying window now. High fit + high intent = call today. High fit + low intent = nurture. Low fit + high intent = trap (they'll take demo time and never close). Low fit + low intent = ignore.
Firmographic: industry, headcount, revenue, geography. Fit-to-use: technology signals (do they use tools that suggest they'd benefit), organizational signals (do they have the roles that use your product), business model signals (subscription vs transactional). Score each dimension 0-3, weight by predictive power (from closed-won analysis), sum to a fit tier (A/B/C/D). Update the weights quarterly based on actual close rates by tier.
Real intent signals: pricing page visits, demo requests, multi-page product tours, comparison content ('vs Competitor'), pricing calculator use, live chat inquiries, second-visit within 7 days. Weak signals often mistaken for intent: email opens, LinkedIn ad clicks, blog reads, whitepaper downloads. Weight the strong signals 5-10x the weak ones. A lead who visited the pricing page twice this week outranks one who downloaded 5 whitepapers over a year.
Vendors like 6sense, Bombora, and G2 provide 'anonymous' account-level intent (research activity across the web tied to companies, not individuals). Useful for account-based motions: 'Company X is researching CRM alternatives — reach out.' Less useful for individual lead prioritization. Cost: $50-150K/year for enterprise-grade data. Worth it above $50K ACV with account-based sales; overkill below.
MQL (marketing qualified lead): passed fit threshold, some intent. SQL (sales qualified lead): SDR-validated fit and intent, worthy of AE time. Opportunity: AE-validated need, budget, timeline. Each stage has explicit criteria, and each conversion rate is measured. If MQL→SQL is under 30%, marketing's definition is too loose. If SQL→opportunity is under 40%, SDR qualification is too shallow.
Complex point systems no one trusts: reps ignore scores that don't match their instincts. Not calibrating with closed-won data: scores predict something, but not the thing that matters. Rewarding email opens: opens are noise, not intent. No fit filter: high-intent low-fit leads eat rep capacity. Not updating weights: what predicted closes 2 years ago doesn't predict them today.
Investor directory · Fundraising library · Articles A–Z · Company funding database