Raw user feedback from three public channels, grouped into
distinct pain points and ranked by how many independent sources report each
one. Generated by a pipeline, interpreted by a human.
293
complaints analyzed
3
data sources
6
pain points found
281
praise filtered out
Prioritization lens: Improve retention among free-tier and lapsed users
Rank by
How this works
Collect — reviews and posts are pulled from public feeds and saved to a growing dataset.
Filter — pure praise is dropped. This is pain-point discovery, so a five-star "love it" carries no signal.
Group — feedback that shares distinctive vocabulary gets grouped together. No topics are defined in advance; the groupings emerge from the language itself.
Rank — by default, by how many independent channels report the same problem. A complaint three channels agree on is more trustworthy than one loud channel.
Why source agreement beats raw volume. App Store reviewers and Reddit
power users complain about almost entirely different things. Casual users are
loud about ads and skip limits; power users are loud about AI-generated music
and app slowness. Ranking purely by mention count would just surface whichever
channel happened to be noisiest that week.
What this deliberately does not do
Decide anything. It surfaces and ranks. Choosing which problem to
pursue, and defending why not the others, stays a human judgment call.
Size in dollars. No revenue or usage data is wired in. Any synthetic
figures added later will be labelled as such wherever they appear.
Name the problems for you. Cluster labels are the keywords that
define each group. Turning those into real problem statements is the
interpretation step, and it belongs to the PM.