Imo best approach is to make a list of all the questions you’ll have “Eg what % of users make it past each step”, then ask an engineer how much effort it will take to implement the logging for each one, and then prioritize them based on effort vs impact, and let the engineer figure out how to instrument the event calls
I totally agree with that philosophy, and I’ve personally had success with that approach working with teams with more established product analytics.
This particular case is such a bear because we’re instrumenting one sliver of an existing app for the first time. We don’t have prioritization/trade-off decisions to make at this stage, as actual implementation effort is pretty light. The cognitive load right now is really more in the data architecture/engineering.
So either my brain will be melting or an Engineer’s brain will be melting.