3 Comments
User's avatar
Ernle's avatar

A/B testing works in the same sense that greedy methods work in algorithms. At each point, there is a test and you make the best local choice. But surprise, making the best local choice doesn't always lead to the overall best result.

aside for Trung: https://www.wsj.com/business/retail/trying-to-dethrone-the-birkin-make-fewer-bags-2ddf733d is relevant to one of your other interests.

Expand full comment
Trungphan2's avatar

Thanks for the read Ernie! Very true about unintended consequences...Frisby says in his chat that they have counter metrics to make sure things dont go awry (but didn't mention any more details).

Queue-ing up that article now. Very relevant as my wife has been mentioning the Row a lot recently!!!

Expand full comment
Teddy21btc's avatar

A/B testing is evolutionary.

Expand full comment