Not sure what to analyze first?
Choose the symptom that feels closest. After the third answer, you’ll get one analysis to check first.
1/3
Choose one answer to continue.
Before you diagnose
Decompose cost and installs by channel and campaign to separate volume, mix, and efficiency effects.
No. Separate natural trend and seasonality, then validate with a holdout when possible.
No. It proposes hypotheses and a check order while separating observational association from causal evidence.