Aniket,
I think you should definitely already have an idea about what your North Star Metric (i.e. the engagement metric encapsulating the true value of your product) might be before you implement tracking, yes.
However, from a *very* early stage, you will want to see not only this number improving (hopefully), but also which users are responsible for this movement (basic attribution / segmentation), how new cohorts of users are affecting this metric vs the older ones and — crucially—which users are not contributing to this increase and where are they dropping off (Funnel Analysis).
In addition, what if you’re completely wrong about what users will love about your product? If you are only tracking your North Star, but actually users are going wild for feature X — which you thought was a sideshow—you might be missing crucial information that would inform future decision-making.
Consider another case: users *might* love the feature which you’ve based your North Star metric around, but are failing to reach it because of signup or other friction… tracking other events and funnel analysis would show you what’s happening there.
At an early stage, there are lots of unknowns… it pays to be well-informed, though there are diminishing returns. The challenge for a launch-stage team is finding balance between data and action, which is why it’s so tough to determine what level of tracking is minimum *and* viable.
