Reading Data as Business Evidence

This chapter focuses on the three reading errors that sit behind almost every analytical mistake a manager hears about — double-counted revenue, joins that explode, a 4.2-star average that misleads, a quarterly ranking that rewards a fading store. Working through a single week, then eight weeks, at Bean & Basket Coffee, it shows how the grain of a row, the shape of a table, and the measurement type of a column quietly set the ceiling on what any analysis can honestly claim. The takeaway is a posture rather than a formula: read the panel before you rank, keep the finest grain you can sustain, and write the one-page variable dictionary that catches type confusion at design time instead of slide-review time.