The car industry A/B tested selling a car with and without CarPlay

Or drivers prefer buying a Honda. They might not notice that they are the same cars. They should have interviewed a sample of the buyers but they are probably unable to do it.

And then there is the matter of the price. The links in the post point to a $39,900 Honda and to a $44,700 Chevrolet. I'm not surprised that the cheaper car sold more.

I do believe the choice helps. But i also believe Honda has better brand appeal overall and so it's not a/b testing as much as "shiny a++" / "dull commoditized b-" testing. It wasn't about carplay, it was about brand affinity preferences.