An analyst claims that a marketing campaign caused growth solely because both rose together in observational data. Which approach best fits?

1. Define a governed shared semantic model and approved business metrics.
2. Separate correlation from causation and evaluate a suitable experimental or causal design.
3. Normalize currency consistently and document the conversion basis.
4. Use a time-series visualization with a clear time grain.