An organization wants to use multiple marketing datasets to forecast user acquisition. How should they use cloud technology to gain new insights from the data?

Importing into a data warehouse and archiving old data organizes storage but does not generate new insights or produce forecasts from combined sources.

Selectively archiving data in a data lake reduces the volume of data available for analysis, which limits rather than enhances predictive modeling.

Separating datasets and running ML on each independently prevents the model from learning cross-dataset correlations, reducing forecasting quality.

Machine learning models improve in accuracy and generalization when trained on larger, more diverse datasets, so combining multiple marketing sources maximizes the signal available for user acquisition forecasting. Unified datasets allow the model to detect cross-channel correlations that would be invisible if datasets were analyzed in isolation. This approach leverages the cloud's ability to ingest and process large-scale heterogeneous data for ML workloads.