ReviewFold
Guide · Getting the data out
Export App Store and Google Play reviews to CSV
You already have an app and you want the reviews in a spreadsheet, not in someone else’s dashboard. This page shows the seven columns the export produces, one row as it actually arrives, and the encoding details that decide whether the file opens cleanly or turns your Spanish reviews into question marks.
01 · Before you start
Three limits, stated before the steps.
These are the reasons the export looks smaller than a scraper. They are also the reasons it is defensible if someone asks where the numbers came from.
- One app, one region.
Each report covers a single app in a single storefront. Two storefronts mean two reports and two files.
- Up to 200 recent reviews.
That is the sample the report was built from. The file contains that sample, not the app’s review history.
- Paid report feature.
The CSV is exported from a completed report, so it needs full report access. The free preview stops at themes and counts.
If what you need is a place to record your own coding decisions rather than the raw reviews, the 13-column worksheet is the closer tool. This page is about the reviews themselves.
02 · The path to the button
Five steps, using the labels on screen.
The export lives inside the report rather than in a settings menu, because a file has to belong to a sample someone can point at.
- 01
Paste the store link
On the home page, paste the app’s Apple App Store or Google Play URL. The app ID and store are read from the link.
- 02
Pick the storefront and the sample size
Choose the region, then 20, 50, 100 or 200 recent reviews. The sample size becomes the row count in the file.
- 03
Generate the report
The grouping has to exist before there is a report version to export from. This is the step the paid report covers.
- 04
Open Source reviews
In the report, switch from the Report tab to Source reviews. That tab lists the reviews behind the themes.
- 05
Choose Export all reviews (.csv)
The button sits at the top right of that tab. The file downloads with a name built from the app, store, region and sample.
“All reviews” means all the reviews in this sample — the same 200 rows you can page through on screen. It does not mean the app’s full review history.
03 · The seven columns
What each column holds, in the order it appears.
Order matters if you are merging files or writing a formula. The header row is exactly these seven names.
| Column | What it holds | What to watch |
|---|---|---|
| Review ID | The store’s own identifier for that review. | Stable per review. Use it to de-duplicate and to point back to a row after you rename a theme. |
| Date | The timestamp the store reports, in ISO 8601 and UTC. | Not the date you collected it. Sorting as text works because the format is fixed-width. |
| Rating | A whole number from 1 to 5. | Published as left. Nothing here averages or reinterprets it. |
| Title | The review’s headline, exactly as written. | Typos and all caps are part of the text. Do not clean them before quoting. |
| Review | The body text of the review. | Check missing text against the saved source. This export is a sample of written reviews, not the store’s full set of ratings. |
| App version | The version the store reports for that review, when it reports one. | Often empty: the Apple public-web sample we saved did not carry a version. The cell stays blank rather than being filled by guesswork. |
| Source URL | Where the sample came from for that store. | Useful for keeping a citation next to a row you quote. It is a feed URL, not a permanent per-review permalink. |
Two rows in a file can share a rating and a date and still describe different tasks. That is why the export keeps the review text and not just the numbers.
04 · One real row
One saved review, in the exporter’s column order.
Taken from a saved September 2026 sample of 200 Threads reviews on the US App Store. The review values are retained, App version is empty, and the source URL is shortened for display.
Six of the seven cells are short; the URL is the reason the file reads badly on a narrow screen. The row above is not padded to fit — the empty cell is empty in the file too. Long parameter strings are trimmed here with the app and region left visible.
05 · What the file handles
Four decisions that stop a spreadsheet from mangling reviews.
None of these are visible when the file opens correctly. All of them are visible when it does not.
- UTF-8 with a byte-order mark
Helps spreadsheet applications recognise multilingual text. If characters display incorrectly, import the file explicitly as UTF-8.
- CRLF line endings
CSV records use CRLF separators. A quoted review may contain its own line breaks, so physical text lines are not a reliable review count.
- Every cell quoted
Commas, quotes and line breaks inside a review stay inside that review’s cell. A quote inside the text is doubled, which is what the format requires.
- A guard on formula starts
A cell beginning with =, +, - or @ gets a leading apostrophe, so a review that starts with “=” is read as text rather than executed as a formula.
The trade-off is deliberate: the apostrophe is visible in the cell. A review that opens with a dash loses nothing of its meaning, and the prefix helps prevent spreadsheet formula interpretation; confirm how your spreadsheet imports the file.
06 · Opening it
Three ways in, one fallback.
If the text looks wrong, the problem is almost always the import step rather than the file.
- Excel
Double-click. If a column splits in the wrong place or accents break, use Data → From Text/CSV and choose UTF-8 with comma as the delimiter instead.
- Google Sheets
File → Import → Upload, then keep “Comma” as the separator. Importing rather than dragging also preserves long IDs as text.
- Numbers
Double-click, then check that Review ID is text rather than a number, so leading digits are not reformatted.
We have not tested every version of every spreadsheet application. The file format is the ordinary one; if a specific version misreads it, the import dialog is the place to fix it.
Boundaries
What this export will never be.
A file that does one thing can be trusted about that one thing.
- Not a free bulk downloadThe CSV comes out of a completed report, after the grouping exists.
- Not the full review historyIt is the 200-review sample, with the collection date attached.
- Not a collectorThere is no scraper here and no scheduled export. One report, one snapshot of one app in one region.
- Not sentiment-labelledNo score, no theme column, no ranking. Those belong to the report, not the raw file.
07 · Working with it
The file is the start of the work, not the result.
A spreadsheet of raw reviews is a pile of sentences. What turns it into research is a second layer: for each review you keep, what was the person trying to do, what blocked it, and what would you need to see to believe it.
That second layer is what the 13-column worksheet is for. Keep the export beside it, hold on to the Review ID, and every conclusion you write can be walked back to a row.
If what you want is the grouping done for you first, with excerpts attached so you can overrule it, that is what the report does — the export is how you take its evidence with you afterwards.
Current support, stated plainly
- Apple App Store: US and Canada. Google Play: US.
- Up to 200 recent public reviews per report.
- One app and one region per report; no cross-app comparison.
- CSV, PDF and review excerpts are part of the paid report.
- No scheduled exports, no full-history download, no free bulk export.
Get the rows
Analyse an app, then take the reviews with you.
One app, one region, up to 200 recent reviews. The report groups them and keeps the originals attached; the CSV hands you the same rows in a spreadsheet.