Definition:
Multi-armed bandit pricing is an algorithmic approach that continuously tests multiple price options while gradually shifting traffic toward the prices that perform best. The system balances experimentation with performance optimization to identify profitable price points faster.
Benefits:
- Accelerates discovery of high-performing prices.
- Reduces revenue loss from prolonged testing.
- Continuously improves pricing decisions as new data arrives.
Example:
A Shopify merchant selling a backpack tests prices of $45, $49, and $54. The system initially distributes traffic across all three prices to learn how customers respond. As performance data shows that $49 generates the highest profit per visitor, the algorithm gradually allocates more traffic to that price while still occasionally testing the others.