How nsight turns 150 reviews into 5 priorities.
A look inside the analysis pipeline: why 150 reviews, what the AI produces, and how issues get ranked into an action plan.
Why 150 reviews
nsight analyzes up to 150 of the most recent public reviews for a given app. That number is deliberate:
- Recency matters. Old reviews reflect old versions. The most recent 150 give you a current-state picture.
- Volume matters. 150 reviews provide enough data for patterns to emerge statistically. A single angry review is noise. Thirty reviews mentioning the same crash is signal.
- Public data only. nsight reads what is already publicly visible on the App Store and Google Play. No SDK, no integration, no access to private user data.
Reviews are cached for 7 days. If you generate a second report within that window, it uses the same dataset. After 7 days, a fresh pull picks up new reviews.
What the AI produces
Each report contains a structured set of outputs, all derived from the review text and metadata:
Sentiment score
A 0–100 score with trend direction (improving, stable, declining) and risk level (low, medium, high).
Category scores
Performance, usability, features, and stability each scored individually on 0–100.
Personas
4–8 user archetypes built from the actual review language, with descriptions and representative quotes.
Pros and cons
A summary of what users praise and what they criticize, grouped by theme.
How prioritization works
The action plan is the most directly useful part of the report. It lists the issues identified in the reviews, ranked by a combination of two factors:
- Frequency — how many reviews mention the issue. An issue that appears in 40 of 150 reviews is more pressing than one in 3.
- Severity — how badly the issue affects the user experience. A crash is more severe than a cosmetic complaint.
The result is a ranked list where the top items represent the highest-impact opportunities. Fix the top of the list first, and you address the issues that affect the most users the most severely.
Critical issues are also flagged separately with severity labels, so items that are rare but catastrophic (data loss, security concerns) do not get buried.
Sharing with the team
Every report gets a shareable link. Anyone with the link can view the full report without needing an nsight account. This means you can paste it into a Slack channel, a Jira ticket, or a planning doc and everyone sees the same evidence.
On the Pro plan, PDF export is available for teams that need offline or archival copies.
Try it
The free tier includes 2 reports per month. Search for any public app, generate a report, and see the full output, including the prioritized action plan. No credit card required.
See what your reviews are saying
Generate your first feedback intelligence report free, no credit card required.
Get started free