Frequently asked questions
What's the simplest scoring model that actually works?
A weighted sum of fit (firmographic match to ICP), intent (buyer-intent signals in the last 14 days), and trigger events (funding, hiring, leadership change). Start with three inputs and equal weights, then tune from won-deal data.
Should account scoring use machine learning?
Only if you have at least 200 closed-won deals to train on. Below that, a transparent rules-based model beats ML because it's auditable, explainable to AEs, and easier to debug when scores look wrong.
How do you keep account scores from going stale?
Rebuild the model quarterly using the most recent four quarters of won/lost data. Score everything else daily off the live model. The model itself should change rarely; the inputs feed it continuously.
What's the right tier breakdown for a scored list?
Top 10 to 20 percent = Tier A (worked aggressively), next 20 to 30 percent = Tier B (lighter touch), rest = Tier C (newsletter, retargeting). The exact percentages depend on SDR capacity; the principle is concentration.
What does Landbase score on by default?
Firmographic fit, technographic match, hiring velocity, funding recency, intent overlap, and decision-maker tenure. The weights are tunable through the CLI so RevOps teams can match the model to their own won-deal patterns.
.png)
.png)