As soon as the company starts collecting metrics and sets up A/B testing infrastructure people can just implement whatever they want, as long as they keep impro

As soon as the company starts collecting metrics and sets up A/B testing infrastructure people can just implement whatever they want, as long as they keep improving these metrics. Very often there's no one who can take a look at the resulting mess and decide enough is enough. It's all just self-congratulatory meetings with extremely optimistic charts and an occassional employee complaining about the product being utterly broken and unusable.