Subscription Refunds Reporting: Where to Find and Read Them

Subscription Refunds Reporting: Where to Find and Read Them

Subscription refunds are reported as their own event in each store, and they land in the month the refund is processed, not the month of the original sale. To report them correctly, pull refunds from each store’s native report, keep them separate from cancellations, and view them next to gross revenue and subscriber counts in one place. This guide shows where each store puts refunds, how to read the rows, and how to build a refund rate you can trust across iOS and Android.
TL;DR
- A refund is not a cancellation. Apple and Google both count them separately: a cancellation stops future renewals, a refund returns money already paid.
- Apple flags refunds with a Refund field in the Subscriber Report and shows negative units and customer price in the Summary Sales Report. Google Play shows Refunds and Partial refunds columns in Financial reports.
- Refunds hit the period they are processed in, so a January purchase refunded in February lowers February revenue.
- Calculate refund rate with one fixed definition on both platforms, then watch it by product, country and cohort.
- Seeing refunds beside downloads, revenue and subscriptions in one dashboard is what turns a stray negative number into a signal.
Table of Contents
- What a refund means in subscription reporting
- Where to find refunds in each store
- Quick reference: how each report shows a refund
- How a refund changes your numbers
- Does canceling create a refund?
- Reconciling refunds: the five checks that find most gaps
- How to calculate a refund rate you can trust
- Turning refund data into decisions
- Why one dashboard beats three exports
- Perspective
- Start with the data you already have
- FAQ
- Sources
What a refund means in subscription reporting
A refund is money returned to a subscriber for a charge that already settled. That sounds obvious, but the reports blur it with three neighbors:
- Cancellation: the subscriber turns off auto-renew. Access usually continues to the end of the paid period. Apple’s subscription Events view counts a cancellation when auto-renew is turned off, or when billing retry fails at the last attempt, and counts refunds as a separate figure that includes full and partial refunds.
- Refund: the store returns money. Apple’s subscription events reference defines two refund event types: “Refund” and “Refund After Renewal Extension”.
- Chargeback or dispute: the cardholder’s bank reverses the charge. For store purchases this is handled between the store and the payer, so you mostly see it as a refund row. If you also bill on the web, disputes appear in your billing tool instead.
- Billing reversal or adjustment: a correction to an earlier figure. Amazon’s Reporting FAQ explains that an adjustment can lower previously reported sales, or raise them when a refund did not go through.
The practical rule: never treat a canceled subscription as lost revenue and never treat a refund as churn. They answer different questions. Cancellation tells you about retention. Refunds tell you about expectation gaps, accidental purchases, pricing confusion and support friction.
Where to find refunds in each store
Apple App Store Connect
Apple gives you three useful views.
- Subscription Events. This is the quickest check: a count of refunds per period next to cancellations and renewals.
- Subscriber Report. Apple’s Subscriber Report reference says a Refund field equals “Yes” for full or partial refunds, refund rows carry negative Customer Price values, and the Purchase Date field holds the original purchase date. A Units value of 0 may indicate a partial refund, and a refund with a positive customer price implies a refund reversal.
- Summary Sales Report. Per Apple’s Summary Sales Report reference, refunds appear as negative Units and Customer Price but positive Developer Proceeds, and the report separates Customer Currency (the customer’s three-letter ISO currency code) from Currency of Proceeds (the currency you are paid in).
That sign convention trips up spreadsheets. If you sum Units and Developer Proceeds without checking the sign logic, refunds can quietly inflate proceeds in a pivot. Decide once how you will flip signs, and document it.
Google Play Console
Google’s subscription overview under Financial reports shows active, new and canceled subscriptions plus a Refunds metric, which is the total number of refunds issued in the selected period. Gross revenue there is what users paid, including taxes and other fees.
For money-level detail, the revenue page shows per-product revenue with separate Refunds and Partial refunds columns. Its financial data is based on the UTC time zone, refreshed daily, and derived from Estimated Sales Reports that you can export from the Financial reports page. Those reports are cumulative for the current payout period and are updated nightly, adding transactions as accounts are charged or refunded, according to the Android Developers Blog. Transactions can take up to about two days to appear, depending on time zone differences.
Other stores and billing systems
If you also publish on Amazon, remember its rule for 2018 and later: an adjustment or refund is reported in the month it is made or issued, not in the month of the sale (Amazon Reporting FAQ). The same pattern holds across stores, so refund timing is the first thing to check whenever two reports disagree.
Quick reference: how each report shows a refund
| Report | Where the refund shows | What to watch |
|---|---|---|
| App Store Connect Subscription Events | Refunds count, separate from cancellations | Includes full and partial refunds |
| Apple Subscriber Report | Refund field = Yes, negative Customer Price | Units of 0 may mean a partial refund; Purchase Date is the original date |
| Apple Summary Sales Report | Negative Units and Customer Price, positive Developer Proceeds | Customer Currency differs from Currency of Proceeds |
| Google Play subscription overview | Refunds metric for the period | Gross revenue includes taxes and fees |
| Google Play revenue page | Refunds and Partial refunds columns per product | UTC time zone, refreshed daily |
How a refund changes your numbers
A refund touches four figures, and each moves differently.

Revenue. In accounting terms, a refund is not just a negative line. Stripe’s revenue recognition methodology explains that cash goes back to the customer, previously recognized revenue is offset through a contra revenue account, and unrecognized deferred revenue is cleared. With its option to record refunds separately, a 90 USD payment fully refunded in the same period shows 90 USD revenue and 90 USD refunds, for a net of 0 USD.
Proceeds and fees. The platform fee on a refunded charge is returned too. Chargebee’s Google Play documentation says a charge refund row produces a credit note, while a Google fee refund row reduces the expense tied to that refund. Your net loss on a refund is therefore smaller than the gross amount, so compare gross to gross and net to net.
Partial service already delivered. Stripe’s Apple App Store connector example walks through a three-month subscription bought January 2 for 91 USD (90 USD developer proceeds) and fully refunded February 1. Thirty days of service were used, so 30 USD stays in the refunds balance, and the remaining 60 USD of deferred revenue is cleared.
Subscriber counts. Apple’s Events view tracks refunds apart from cancellations, so a refunded subscriber can still appear in some cancellation or active figures depending on timing. Reconcile counts at the transaction level rather than trusting a single dashboard tile.
Does canceling create a refund?
No. Canceling turns off renewal. It does not return money by itself, and access normally continues to the end of the paid term. A refund is a separate action approved by the store, or by you if you bill directly.
When a store does grant a full refund on a canceled subscription, accounting tools treat it as a reversal. Chargebee’s Apple App Store documentation states that its revenue recognition reverses cumulative recognized revenue, zeros out deferred revenue, processes the refund as a credit note based on the service start and end dates, and includes the platform charge reversal in the current open period.
For reporting, that means two different dates matter: the original purchase date and the refund processing date. Keep both in your data.

Reconciling refunds: the five checks that find most gaps
When your store report, analytics and finance numbers do not match, work through these in order.
- Check the period. A refund is deducted in the month it is processed, not the month of the sale. A refund for a January renewal that appears in February is not an error. It is the most common cause of mismatches.
- Check the sign. Apple’s Summary Sales Report shows negative units but positive Developer Proceeds on refund rows. Normalize signs before summing.
- Check the currency. Match Customer Currency to Currency of Proceeds. A refund in a local currency is converted into your settlement currency, so the amount you see in your payout can differ from the amount the customer paid.
- Check partial versus full. Apple uses a Units value of 0 as a hint for partial refunds, and Google splits Refunds from Partial refunds. If you count rows as refund events, partial ones can inflate your count relative to money.
- Check the source of truth for billing. If you also bill outside the stores, you need each billing system’s own adjustment data. Paddle, for example, documents each adjustment’s earnings in the developer’s balance currency, as the total minus the fee, in its adjustment reports.
Stripe’s team also recommends linking each refund to its original purchase and adjusting deferred revenue, instead of treating refunds as plain negative line items, and importing store data daily to reduce month-end corrections (Stripe documentation). The same principle applies if you work in a spreadsheet: add a column for original purchase date so you can attribute each refund to its cohort.
How to calculate a refund rate you can trust
Refund reporting only becomes useful when it is a rate. Pick one definition and use it on every platform:
- Refund rate by count = refunded transactions in a period divided by paid transactions in the same period.
- Refund rate by value = refunded gross amount divided by gross billings in the same period.
- Cohort refund rate = refunds attributed to purchases from one week or month, divided by purchases in that cohort. This is the cleanest view because it avoids the timing mismatch described above.
A worked example. In March you record 2,000 paid subscription transactions on iOS and 1,500 on Android, and 60 and 30 refunds are processed in March. Count-based refund rates are 3.0% and 2.0%. If 15 of the iOS refunds relate to February purchases, your March cohort refund rate on iOS is lower at first and will climb as late refunds arrive, so read cohorts after a fixed window, such as 30 days after purchase.
Then segment it:
- By product and price point, to see whether an annual plan draws more regret than a monthly one.
- By country, because pricing, currency and local expectations differ.
- By acquisition source, to see whether a campaign or a keyword brings in users who ask for refunds.
- By app version, to catch a release that broke a promised feature.
The last two are where refund data meets marketing. A spike tied to one ad network or one store listing change is a listing or targeting problem, not a billing problem. If your screenshots or description promise something the paywall does not deliver, refunds follow.
Turning refund data into decisions
Once the numbers are consistent, each pattern points at a fix:
| Pattern | Likely cause | What to do |
|---|---|---|
| Refunds cluster in the first days after purchase | Unclear trial or price on the paywall | Make price and renewal terms plain at the point of purchase |
| One country spikes | Local pricing or currency mismatch | Review regional price tiers |
| One campaign spikes | Audience or promise mismatch | Align ad creative and store listing with the real product |
| Refunds follow a release | A bug or removed feature | Check release notes and crash data for that version |
| Annual plans refund more than monthly | Sticker shock at renewal | Add a renewal reminder and clarify the plan in-app |
For the listing side of this work, your store metadata must describe what the subscription delivers within the rules of each store, and ASO reporting metrics shows which listing metrics to read alongside revenue. If you run SDK-based subscription tracking, check your Firebase subscription tracking setup so refund-affected subscribers are not double counted.
Why one dashboard beats three exports
The hard part of refund reporting is not finding the number. It is that the number lives in a store console, a billing tool and an analytics tool, each with its own dates, signs and currencies. Every manual export adds a chance to misalign a period or flip a sign.
Apptenium is built around that fragmentation. It brings ASO scanning, keyword tracking and competitor intelligence together with app performance reporting for downloads, revenue, subscriptions and ad monetization, and it integrates data from Firebase, Google Analytics and ad networks. When your revenue and subscription figures sit beside your keyword rankings and listing changes, a jump in refunds can be read against what changed in the listing, the campaign or the release, rather than argued across three spreadsheets.
Perspective
Most refund conversations we hear start in the wrong place: with the refund. By the time a refund shows up in a report, the cause happened days or weeks earlier, in a store listing, an ad, a paywall or a release.
That is why we built Apptenium to keep listing work, keyword tracking and app performance in one platform. Our view is that refunds are a marketing signal as much as a finance one. If you can see downloads, revenue and subscriptions next to the listing changes and campaigns that preceded them, you stop guessing which promise did not match the product.
Our advice: define your refund rate once, track it by cohort, and review it in the same place you review rankings and installs. The teams that do this fix causes instead of chasing totals.
Mike
Start with the data you already have
If you want refund trends beside your rankings and installs, Apptenium has a Free plan with a limited number of scans per month, and a Pro plan at $9.99 per month with unlimited scans and premium features. You can compare them on the pricing page.
If you manage several apps, the cross-app performance reporting guide shows how to prove the effect of ASO work across a portfolio, and the BigQuery app analytics guide covers warehouse-level analysis for teams that want raw data.
FAQ
Is a subscription refund the same as a cancellation in reports?
No. A cancellation records that auto-renew was turned off, or that billing retry failed. A refund records money returned. Apple counts them as separate figures in its subscription Events view, and Google Play shows canceled subscriptions and Refunds as separate metrics.
Why does my refund show up in a different month than the purchase?
Stores report a refund in the period it is processed. Amazon states this explicitly for 2018 and later, and the same logic applies to Apple and Google reporting, so a refund can reduce revenue in a month after the sale. Use the original purchase date to assign refunds to cohorts.
Why are refunds negative in some columns and positive in others?
In Apple’s Summary Sales Report, refunds show negative Units and Customer Price but positive Developer Proceeds. The sign convention differs by field, so normalize signs before you total anything.
How do I know whether a refund was partial?
In Apple’s Subscriber Report, the Refund field says Yes for full or partial refunds, and a Units value of 0 may indicate a partial one. Google Play separates Refunds and Partial refunds in its revenue page columns.
Do platform fees come back on a refund?
For Google Play, Chargebee’s documentation describes a Google fee refund row that reduces the expense tied to the refund, so the fee is reversed as well. Track gross and net refund amounts separately to avoid overstating the loss.
Sources
- Apple: Subscription events reference
- Apple: View subscription data
- Apple: Subscriber Report reference
- Apple: Summary Sales Report reference
- Google Play Console Help: Review in-app subscription performance
- Google Play Console Help: Review your app’s revenue and buyer data
- Android Developers Blog: New Sales Reports on Google Play
- Amazon Developer: Reporting FAQ
- Stripe: Revenue Recognition with refunds and disputes
- Stripe: Connector for the Apple App Store
- Chargebee: Google Play Store RevRec
- Chargebee: Apple App Store RevRec
- Paddle: Adjustment reports
