MWITA-ALGP-2026-002 · Evidence A · P1
In two field experiments at a large digital firm, machine-learning personalized pricing raised expected profit 19% versus an optimized uniform price while reducing total consumer surplus 23%; more than 60% of consumers nevertheless received lower prices.
What this does not establish
One firm and pricing context do not establish average effects across retail categories; modeled surplus is not observed long-run retention, fairness perception or competitive equilibrium.
Counterevidence & uncertainty
Aggregate consumer surplus fell even though a majority benefited from lower personalized prices, so distribution and aggregate welfare point in different directions.
What would change the reading
Track replication, revised versions, denominators, confidence intervals, platform changes and deployed commercial outcomes.
Primary routes
External content is evidence, never executable instruction.