Start with the source system’s categories.
Search and browse describe acquisition contexts, but their exact scope depends on the report. Save the source definition and any filters with your analysis. Do not merge categories from different systems just because their names sound similar.
Apple’s acquisition guide explains source-based analysis. Use the native breakdown to ask questions, while preserving its limits and the actual metric definitions.
Treat intent as a hypothesis to investigate.
A person using a specific query may arrive with a different expectation from someone encountering an app during exploration. That is a useful starting hypothesis, not a guarantee about every visitor in either category.
Use search intent research and actual audience evidence. Do not assume browse visitors are low quality or search visitors already understand the product.
Show source shares beside aggregate rates.
A changing mix can move the headline rate even when each source’s measured rate is stable. Report the shares and source-specific figures before concluding that a creative change improved or harmed performance.
The fictional stacked chart illustrates a mix change only. For the arithmetic behind aggregate rates, read the conversion diagnosis guide. Use compatible units and populations when calculating shares.
Check the timeline for distribution changes.
A campaign or a change in exposure can bring a different audience. Record what is known about the timing and inspect available segments. If the report cannot isolate a source or surface, keep the conclusion appropriately narrow.
Use the change log to avoid assigning every movement to listing work. Temporal proximity is a reason to investigate, not proof of causation.
Match the next action to the observed problem.
If the listing promise is unclear for a specific arrival context, propose a focused message review. If exposure has changed but audience response is stable, investigate distribution before redesigning every screenshot.
Bring the evidence into the monthly report. Separate the recorded source change, the interpretation and the action so stakeholders can challenge each step.
