|
|
|
|
|
by pjlegato
1120 days ago
|
|
It's even older than that: Amazon began doing this in the 90s. Their initial idea was to synchronize warehouse inventory with the online store in realtime, so that users would never buy anything out of stock. That proved logistically difficult and expensive, so they decided that the frontend would merely checkpoint inventory levels at intervals. When someone inevitably ordered something that was no longer in stock, they simply sent a robo-apology note and refunded the order. Mitigating the hit to customer satisfaction was deemed cheaper than the very expensive proposition of synchronizing distributed warehouse inventory levels in realtime over unreliable networks. |
|