Making decisions based off an any metric is very dangerous and the downfall of many products.
You MUST have a deep understanding of exactly what that metric is measuring and all of the things it doesn't take into a account.
For a concrete simple example. An A/B test may sure that forcing users to create an account before seeing shipping charges increases order completion by 10%. But then it turns out you've reduced return visits to the site by 30% and decreased overall conversion rates due to damaging that funnel.
Reality is very complex. Trying to boil everything down to a few numbers to make decisions based on seems like it simplifies things but often times you are just ignoring the complexities and flying blind by chasing metrics.
You MUST have a deep understanding of exactly what that metric is measuring and all of the things it doesn't take into a account.
For a concrete simple example. An A/B test may sure that forcing users to create an account before seeing shipping charges increases order completion by 10%. But then it turns out you've reduced return visits to the site by 30% and decreased overall conversion rates due to damaging that funnel.
Reality is very complex. Trying to boil everything down to a few numbers to make decisions based on seems like it simplifies things but often times you are just ignoring the complexities and flying blind by chasing metrics.