Statistics you can defend
Primary and secondary goals
One objective decides the experiment. The rest are context, and saying which is which up front is most of what keeps a programme honest.
A website marks which of the events it emits are allowed to be objectives. Newly discovered events start ignored, because the tracking server registers every event name it sees and an opt-in is what keeps the list from filling with whatever the site happens to fire.
From that pool, each experiment picks one primary goal and lists whatever else it wants to watch as secondary. The primary is what the verdict is computed against, what the attention queue reads, and what the automated stopping rule acts on. Everything else is there to catch the win that cost you something elsewhere.
Naming the primary before the experiment starts is the part that matters, and it is why the platform makes you do it rather than letting you choose a metric after the fact. Picking the goal that happens to have moved is how a programme accumulates wins and no revenue.
A subtlety worth knowing: demoting an event on the website afterwards does not change experiments that already chose it. A finished result must not change retroactively because somebody tidied up a mapping months later.