We have a web-based enterprise analytics application, with many potential user interactions spread across many different reports. Thus we have potentially hundreds of trackable events.
We are reviewing a couple of different options for our Google Analytics events tracking.
Option A is to track "everything" -- mainly all potential user clicks, e.g. clicking on a data point in a trend chart, clicking on a filter toggle, identifying the option selected from a filter list, etc.
The engineers would implement this systematically, and we would have an agreed naming convention, whereby the event category, name and label would identify, say, the report page name, event type and user's selection respectively.
Option B is to first identify a list of events that we want to track, say a top 10 list of what might we would agree are the most important metrics , and ask the engineers to track only those things. We could increase this list over time, but this would require queueing the request with the engineers.
We have already identified pros and cons to both approaches but, from a UX research point of view, which option is preferable?