Lead scoring in isolation misses half the picture
A contact score without account context produces high-scoring leads at companies you'd never want to work with.
Traditional lead scoring assigns points based on contact behaviour: opened 3 emails (+10), visited pricing page (+20), downloaded a whitepaper (+5). Hit a threshold, route to sales.
The problem: a contact at a 5-person startup can accumulate the same score as a contact at a 500-person Series C. The rep receives both as “hot leads.” One is a conversation worth having. One is a waste of 45 minutes.
Account-level context has to be a gate, not a modifier.
The scoring model I build has two layers:
Account gate (binary). Before any contact score is considered, the account has to pass a firmographic filter. Industry, employee count, revenue range, tech stack. If the account fails the gate, the contact score is irrelevant, the record doesn’t route to sales regardless of behavioural score. It stays in nurture or gets suppressed entirely.
Contact score (weighted). Only applies to contacts at accounts that passed the gate. Behavioural signals (page visits, email engagement, content downloads) combined with fit signals (seniority, function, decision-making authority). The weighting should reflect what actually correlates with closed-won in your specific pipeline. Pull a sample of your last 50 closed-won deals and work backwards from the signals that were present.
The third dimension worth adding: account engagement score. Not just one contact’s behaviour, but the aggregated signal across everyone at the account. Three different people from the same company visiting your site in a week is a stronger signal than one person visiting three times. HubSpot doesn’t calculate this by default. You need to build it in Clay or n8n.
When all three align (account passes the gate, contact has high fit + behavioural score, account has aggregate engagement) you have a genuinely hot lead worth a rep’s time.