Lead Scoring and Routing for B2B SaaS: A Worked Example

Start with two scores

Use separate tables for account fit and recent intent. All weights and thresholds are illustrative starting assumptions, not validated defaults. Agree your ICP and exclusions with sales before implementing them.

Keep timestamps and source records. In this example, count each event type once within its window, expire it afterwards and cap intent at 100. Repeated page refreshes must not inflate a score. Missing enrichment is unknown, not evidence of poor fit.

Three worked records

Fictional accounts; the sums demonstrate the model, not client results.

Route in a defined order

First deduplicate against the stable CRM record. Respect suppression and consent settings. Existing customers go to their account owner; existing opportunities stay with the opportunity owner. Missing ownership or required data goes to a named exception queue.

For remaining new leads in this illustrative model:

1. Fit below 60: manual qualification if a demo was requested; otherwise appropriate nurture.
2. Fit at least 60 but intent below 60: nurture until a relevant change.
3. Fit and intent at least 60: route by territory, then size. More than 1,000 employees goes to the enterprise priority queue; 501-1,000 to enterprise; 50-500 to mid-market; fewer than 50 to the small-business queue.
4. If size or territory is unknown, or the owner is unavailable, use the exception queue and record the reason. Assign its owner and response expectation before launch.

Test before activating

The worksheet includes boundary cases at 49/50, 500/501 and 1,000/1,001 employees, stale events, duplicate requests and unavailable owners. Run historical records through the rules without sending alerts or changing owners first.

Freeze each record's inputs at the scoring date. Compare later qualification and retention across score bands, showing the cohort size and observation window. Do not use future revenue to construct a historical score. Log false positives and sales overrides; revise weights when evidence supports a change.