Start with coverage, then compare
Check the last successful sync and the dates each metric covers. Manifest compares two adjacent seven-day windows. Partial sums remain visible, but a missing day prevents a claimed week-over-week gain or decline. A sync today can still contain older Apple data.
Find the stage worth investigating
- Impressions and taps
- Use these to review listing visibility and engagement. Tap-through is taps ÷ impressions × 100 on dates where both exist. Unique counts can repeat across report dimensions, so this is a report-based ratio, not a count of distinct people.
- First-time downloads
- Compare acquisition across fully covered weeks. Downloads ÷ impressions is a directional ratio, not a joined user funnel: someone can download without opening your product page.
- Sessions, deletions, and crashes
- These describe reported usage. They do not measure cohort retention, churn, or a crash-free user percentage. Deletions can come from people who downloaded months ago.
Write down your hypothesis, make one listing or product change, and wait for comparable data before deciding whether to keep it.
Know which money number you are reading
Proceeds are Apple’s estimated earnings after applicable taxes and commission. Manifest sums purchases and refunds on their event dates. Subscription proceeds are already included in total proceeds; adding the two tiles would double-count them. These are not bank payouts or accounting statements.
Estimated monthly recurring proceeds model paid active subscriptions using each product’s observed positive proceeds per purchase and billing cadence. Weekly uses 52 periods per year, monthly 12, and yearly 1. Annual run rate is the same unrounded model expressed over a year. For example, one $6 weekly product and one $36 yearly product imply $348 annually, or $29 monthly.
Free trials do not count as paid revenue. Missing prices, unknown durations, or unmodeled paid offers leave the estimate unavailable. Refunds reduce actual proceeds but are excluded from price samples. Prices, territories, offers, tax, and commission changes mean this remains an estimate of developer proceeds, not Apple’s customer-sales MRR.
Check the per-product calculation on the dashboard before using a run rate to plan spending.
Use the archive to make the next decision easier
Review which apps and territories deserve your next hour. Chart downloads, proceeds, impressions, or crashes over the same dates. Export the record or use Copy for AI to bring both numbers and coverage into your own analysis.
Manifest does not measure in-app activation, cohort retention, advertising income, acquisition cost, or customer lifetime value. Those need additional product events or cost data. It also does not provide competitor keyword rankings.
Definitions and further reading
Related
- Sales and Trends vs Analytics: why the same day shows different numbers across Apple's reports.
- Why App Store Connect shows zero downloads: check coverage before trusting a drop.