Totally agreed, this seems like a problem of picking the wrong metrics to target. But it goes beyond the engineering team picking too narrow a subset of latency

Totally agreed, this seems like a problem of picking the wrong metrics to target. But it goes beyond the engineering team picking too narrow a subset of latency to measure. What metrics were the product and organization leadership team targeting that led them to miss this customer feedback?