A north-star metric is a political object. It keeps a company pointed somewhere. It is a poor diagnostic when the middle of a funnel is rotting while the star still rises on mix shift, seasonality, or a definition change nobody put in the changelog.
Stage health is narrower: for each contract, is the conversion in this window moving for reasons we can name? In GB product orgs we often see weekly business reviews that only show the star. The paid stage can decline while new logos from a partnership inflate the top. Everyone leaves the meeting calm.
The Experiment readouts for funnels programme exists for analysts who already run tests but still present lift as if the funnel were one number. We ask for the stage that moved, the stage that did not, and the star as an appendix. That ordering feels rude the first time. It is also how you stop shipping experiments that “win” on a metric that no longer describes the leak.
Keep the star. Put it second. Product Funnel Analytics is the practice of making the middle visible without pretending the company should abandon a shared goal.