All records are fictional editorial exercises, not KitNelo traffic or AI-product test results. Source CSVs use the same English headers in both packs. Keep originals. Answer files are for comparison, not additional inputs.

A=Period, B=Channel, C=Visits, D=Qualified. Baseline rows 2–3, current rows 4–5. Overall rates: SUM(D2:D3)/SUM(C2:C3) and SUM(D4:D5)/SUM(C4:C5). Rates in expected-metrics.csv are fractions: 0.19 means 19%. Both channel rates improve; overall conversion falls because the visit mix shifts. These four records do not establish causation or statistical significance.
