Forge: Marketplace Incentive Intelligence
AI agents test incentive-allocation strategies, find over-funded supply, and recommend the highest-efficiency allocation.
Problem statement
Marketplaces quietly over-spend on incentives: earner bonuses are set by hand and rarely re-fit to how supply actually responds, so money piles up on supply that would show up anyway. Forge reads a marketplace's supply shape, finds the over-funded and under-funded earners, simulates allocation strategies, and recommends the highest-efficiency allocation, with both the product and the value-creation case. The pattern is platform-agnostic; the demo runs on a synthetic dataset.
Run Demo is an analysis of the default sample dataset.
Run Demo (Alternate Dataset) is an analysis of a randomly selected sample dataset from previously seeded synthetic data.
Sample dataset & assumptions
Data: multiple synthetic marketplace data-sets reflecting different supply patterns, thousands of earners each over 26 weeks (with a seasonal swing across the period)
Per earner: GMV, trips completed, hours active, acceptance rate
Tiers (weekly incentive/earner): Casual $12 · Core $45 · Power $95
Elasticity: supply responds along a saturating curve calibrated to a 0.17 to 0.75 labor-supply range, drawn from published rideshare driver estimates
Saturation proximity is measured against each earner's incentive saturation point.