name: pricing-algorithm description: Reference for the buy box optimization pricing algorithm. Use when understanding pricing decisions, debugging price changes, or investigating buy box behavior. Covers the core optimization loop, buy box detection, and edge cases like multiple buy box winners.

Pricing Algorithm Reference

Overview

The pricing engine uses a buy box optimization strategy that continuously adjusts prices to find the optimal ceiling while retaining the buy box. The algorithm oscillates prices to test boundaries.

Core Algorithm

Goal: Maximize price while maintaining buy box ownership.

Logic Flow:

1. Have buy box + 90%+ ownership over 4 hours → Increase price by 1% (test if ceiling can be higher)
2. Have buy box + <90% ownership → Hold current price (stabilize)
3. Lost buy box → Decrease price to match competitor or drop 1%
4. At minimum price without buy box → Cannot compete, hold position

Key Files:

• BuyBoxOptimizationPricingStrategy.cs - Core pricing logic
• AmazonAnyOfferNotificationsReader.cs - Buy box detection from notifications
• RecommendationContextCollector.cs - Context gathering for pricing decisions

Buy Box Detection

We detect buy box ownership by checking if our specific offer has IsBuyBoxWinner = true:

var weHaveBuyBox = sortedOffersByPrice
    .FirstOrDefault(offer => offer.SellerId == BisonSellerId && offer.IsBuyBoxWinner) is not null;

Edge Cases & Clarifications

Multiple offers on a listing can have buy box, thus we have to filter our own offers which has buybox and take highest priced one to determine if we have buy box or not. Then this does not impact the algorithm.

Price Change Triggers