The Clusters Are Not Segments
Decide what you would do from the brief alone, including whether you would change anything at all. Everything below it is available, but the exercise stops working if you open it first.
Marketing asks: "We ran k-means on customer behaviour and got five clusters. Can you confirm these are our customer segments so we can build campaigns around them?" The clusters have been named — "Loyalists", "Bargain Hunters" — from a 2D plot.
Running a silhouette score and a few more values of k, picking the k with the best score, and confirming the segments because the metric improved. The silhouette measures how well the data fits the clustering's own idea of separation; it says nothing about whether the groups differ in any way marketing cares about. The metric adds a number to a story, the campaigns launch against five named groups, and when two of the campaigns perform identically nobody connects that to the fact that the two clusters differed only on an unscaled spend column.
Read this even if you are confident. It is here rather than behind a button because it is the answer most teams actually ship, it passes review, and its cost arrives weeks later when the labels do.