1 min lesson
Point-based vs. model-based
Match each case in "Point-based vs. model-based" to the signal and response that fit it.
Step 1 of 2
Point-based vs. model-based
Both have a place and the interview reward goes to the one whose tradeoff you can articulate rather than the one you happen to prefer.
+20 for Series A-C, +15 for 10+ active devs, -30 for student email.
Transparent: a rep can see exactly why a lead scored 80.
Start here. Stakeholders trust what they can read.
Fit a model on historical won/lost to weight features for you.
Higher ceiling once you have enough labeled outcomes.
Earn it later; an opaque score erodes trust if it's wrong early.
On a small, flat team, a rep who doesn't trust the score will ignore it and work their own list - and now you've built a system nobody uses. A point-based model where every contributing signal is visible buys adoption. You can graduate to a learned model once outcomes are labeled and people already believe the scores.