Playbook: Building a Partner Lead-Quality Scorecard That Survives Audit

Playbook: Building a Partner Lead-Quality Scorecard That Survives Audit

Volume-based partner payouts reward whoever submits the most forms. A quality scorecard reweights toward downstream value. How to construct one without overfitting:

1. Define the outcome variable first. Pick a single leading indicator that correlates with revenue and resolves in weeks, not quarters — e.g. reaching a qualified-opportunity stage. Closed-won is the true target but too slow to score against.

2. Assemble candidate features. Firmographic fit (employee count, ICP industry), engagement depth (demo attended vs. ebook), and partner-context signals (deal-registered vs. cold referral).

3. Run a simple logistic regression, not a black box. Per several revenue-ops practitioners, interpretable coefficients matter more than a few points of AUC because partners will contest opaque scores.

4. Validate out-of-time. Train on quarters 1-6, test on 7-8. In-sample accuracy is theater.

5. Convert to a tiered payout, not a continuous one — partners optimize against thresholds they can see.

Trade-off: a tight scorecard improves average lead value but shrinks volume, and some partners walk. Model the elasticity before launch.

Guard against survivorship bias: scoring only on leads that entered CRM ignores the leads partners chose not to submit because they sensed rejection.

Open questions: how often to retrain (drift is real as ICP shifts), and whether to expose feature weights to partners — transparency aids trust but invites gaming.