I have a question on the metrics in Experiments and on Dashboard. Currently, I use Binary metrics for retention rates (Day1, Day3, and Day7). I use the same query but change conversion delay for different time frame. Then I add all metrics to dashboard under North Star metric. Should I add only one metric (for example Day1 retention) to dashboard if I want to keep track of Daily Active User? Because right now, on dashboard I see the almost the same numbers from different metrics but different in trends. I think what I see makes sense since I am using the same query for the same metrics. But I’m not sure did I misunderstand how metrics and time delay works.

Conversion window really only makes sense in the context of an experiment since you need some sort of activation event where you start counting. We're working on a better solution for this, but right now metrics that rely heavily on conversion windows (like retention metrics) don't work very well as a North Star metric.

understood. in my case, what do you suggest to use in North Star metric as a workaround? My situation is, the business side would like to see if any features/changes affect the DAU. For the experiment level, the retention rates work well in this case. But they want to see the impact on the DAU too. It sounds confusion, at least to me, that a single change can have any noticeable impact on High Level metric such as DAU. But It’s a step I need to get a buy-in from the business.