ReviewFold
Guide · Monitoring
How to monitor app store reviews without mistaking volume for urgency
Most teams call this monitoring and then check the store when something already feels wrong. A routine that survives a busy week fixes three things up front: what you are watching, how often, and what a single sample is allowed to conclude.
01 · Scope of the word
“Monitoring” is three different jobs.
They need different tools and different routines. Choosing the wrong one is why review checking turns into a chore nobody keeps doing.
See and reply
Individual reviews, answered or escalated. Store consoles cover this for your own app, and a reply tool covers it at volume.
Understand patterns
Recurring feedback across a recent sample: what keeps blocking the same task, and whether it is new. This is the job this page is about.
Detect change over time
Trends, alerts, “it got worse after release 4.2.” Requires stored history and volume. Without both, trend claims are guesses.
The usual mistake is a category error: asking Job C of a tool built for Job B. A report is a snapshot you request, so it cannot tell you what was true last month unless you kept last month’s report.
A useful test before you pick a routine
Write the sentence you want to say at the end. If it starts with “this week, users are still blocked by…”, you need Job B. If it starts with “this got worse since…”, you need history, and no amount of weekly reading replaces it.
02 · Cadence
Choose a rhythm you will actually keep.
A routine that collapses on a busy week produces gaps you cannot interpret. Short and boring beats thorough and abandoned.
- Scan the newest one- and two-star reviews only
- Look for one thing: a task that is completely blocked
- Reply or escalate; do not start an analysis
- Request one report on a fixed app and region
- Read 3–5 reviews behind the top two themes
- Update the sheet with what changed and what did not
- Put two comparable samples side by side
- Mark themes that appeared, disappeared, or changed wording
- Pick one investigation for the month
Before comparing samples, check the app, region, time window, collection method, overlapping reviews and theme definitions. Different sample sizes do not prevent you from comparing the share of reviews mentioning a theme. But a change in that share alone does not establish a trend or show how the experience changed for all users.
03 · Limits
A new label is not a new problem.
Grouping renames things between runs. Before treating any change as a signal, check whether the reviews behind it are actually new.
Not evidence of change
- A theme label that did not appear last week
- A higher count in the same theme
- A lower average rating in one week
- One vivid review that matches your suspicion
Worth a look
- Reviews with dates after the last build shipped
- The same task described in different words
- A theme whose reviews contradict each other
- A problem that persists across three weekly samples
Volume moves for reasons unrelated to product quality: a store featuring, a price change, a competitor outage, or a burst of reviews from one region. Treat a jump as a question — “what else changed this week?” — before treating it as a regression.
04 · Evidence
Always read from the theme back to the review.
The theme tells you where to look. The review tells you whether it is worth looking further.
-
01
Pick the one theme you would act on
Not the biggest. The one where a wrong assumption would cost the most, or where the reviews are most likely to be misread.
-
02
Read three to five reviews behind it
Enough to see whether the group is coherent. If they describe two different tasks, split the theme in your notes rather than in the tool.
-
03
Label the blocker in plain words
Blocked, slow, confusing, missing, or paid — coarse labels survive re-reading. Clever taxonomy does not.
05 · Record
One row a week is the whole system.
If the record takes longer than the reading, it will not survive. This sheet is deliberately smaller than a dashboard.
| Week | Sample | Themes to watch | Blocked task | Decision |
|---|---|---|---|---|
| W1 | app · region · n · dates | two names | verb + object | reply / test / ignore |
| W2 | same scope | — | — | — |
| W3 | same scope | — | — | — |
The first row shows what a filled entry looks like; the others are blank. Empty cells mean “not checked” rather than “nothing found.”
Four rules
Volume is not urgency.
These four rules prevent most bad decisions made from review data. None of them requires a tool.
- Check the date before the countA hundred reviews of an old build are history, not a current problem.
- Check the task before the themeTwo themes can describe one blocked task; one theme can hide two.
- Check for the silent majorityPeople who gave up do not review. Absence of complaints is not absence of a problem.
- Check what else changedPrice, featuring, release and regional events move review volume more than quality does.
06 · Stopping
When to check less often.
A routine that never changes frequency becomes ritual. Set the conditions for slowing down in advance.
- Three weeks with no new evidence: move to monthly, and keep the daily one-star scan.
- The same decision every week: the decision is made; stop re-checking until the next release.
- No usable signal at all: the sample is too small or too generic. Fix the scope before adding frequency.
- A release just shipped: go weekly again for two cycles, then decide.
Monitoring is only worth its cost when a possible answer would change what you do next. If nothing in the last three checks changed a decision, the honest move is to check less, not to build a dashboard.
This method is about reading public reviews. For replies, use the store consoles. For alerts and trends you need history a single report does not keep — ReviewFold’s tracking checks one app daily and emails a Monday summary, which replaces the weekly pass but is not a trend chart. The Google Play analysis method covers the reading step in more detail, and the competitor review guide covers what to do when one of those reviews looks like a feature request.
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-region merge.
- A report is a snapshot you request, not a live feed.
- Tracking is separate: one app checked daily, results on the tracking page and a Monday email. It does not use your monthly reports.
- Tracking sorts new reviews into themes it has already found; it does not score severity or show whether a problem is getting worse.
- No keyword tracking and no reply workflow.
- Free preview shows themes and counts; the paid report adds excerpts, PDF and CSV.
Start with one week
Run one weekly check before designing a system.
Pick one app, one region, one theme you would act on. If that single cycle tells you something you did not know, the routine is worth keeping.