[MetaWhirledPeas]

This is how stats lie. Success is based on measurements across short time spans because that's the only way to reliably correlate features to behaviors. There's no good way to measure the impact of a bad decision across years or decades, so the dark pattern gets away scot free. The only thing stopping it would be a particularly impactful survey or focus group, or worse yet user outcry.

[nine_k]

Most companies do not exist for decades, and many, even for years. They either get acquired, or, more often, dissolved.

Hence the idea to make a quick buck now, like there is no tomorrow, because tomorrow may actually not be %)

[MetaWhirledPeas]

Bingo.

[dnadler]

Not quite sure what you mean, wouldn't an A/B test be able to answer questions like these pretty easily?

[MetaWhirledPeas]

Across what time scale? Are you going to run your A/B test over the course of years to see if the changes drive users away?

[texxtxxet]

This was my argument at the time: build long-term trust with your users (including in your UI) is more important than short-term gains.