This smells like the prototypical A/B-tested product 'optimization' that has become widespread these days. Sure the data probably makes it look like an unequivocal win in the short term, but it compounds ill will towards the platform every time someone goes downstairs to meet their car 10 minutes too early in a downpour.

This reminds me of Schlitz beer and its downfall [0]. Schlitz went from battling Budweiser for domestic supremacy to a case study in how to ruin your brand by repeatedly cheapening your product coupled with a belief that your customers are stupid.

Uber has been like Teflon so far, and Lyft is no Budweiser, but do I wonder if there's a tipping point.

[0] http://www.philcooke.com/the-schlitz-mistake-and-what-beer-c...

I think it's the exact opposite of "users don't notice". I think it's more that people are annoyed but Uber can't measure that easily, so they don't realize. Anecdote: my friends and I have frequently experienced this recently and the result has just been frustration - "it said 3 minutes and now it says 10 minutes" or "why's he driving over there?!". I hang on for the Uber, but if there was a taxi driving by I probably would have hailed it.