Methodology
This page explains, in detail, how ZUT collects, processes, and presents community data. It is a technical companion to our About page. If a number appears anywhere on ZUT, this page explains where it came from and how it was produced.
Where Our Data Comes From
We analyze public discussions on Reddit, Quora, XDA Developers, and MacRumors Forums. For Android devices, we draw on Reddit, Quora, and XDA Developers. For Apple devices, we draw on Reddit, Quora, and MacRumors Forums. When a page compares an Apple device with an Android device, we draw on all four sources.
We do not access private groups, paid forums, or any content that is not publicly visible. We do not collect personal information about the people behind the discussions we analyze.
The Minimum Evidence Threshold
Not every phone has enough public discussion to produce a reliable signal. Before publishing any Community Signal, Community Insight, or related page, we require a meaningful volume of relevant discussion for that specific device or topic.
Where the available discussion is too limited to support a signal, we do not publish an invented number. Instead, the relevant section is omitted, or the page states that community data is currently insufficient.
How We Calculate Community Signals
Community Signals are feature-level satisfaction percentages — for example, how positively owners discuss a phone’s camera, battery, charging speed, or software. Each percentage is derived by analyzing the balance of positive, neutral, and negative sentiment expressed across relevant discussions for that specific feature.
We do not force every phone to display the same set of features. A feature is only shown when there is enough community discussion about it to produce a meaningful signal. If owners rarely discuss a phone’s speakers, for example, that feature is left out rather than filled in with a guess.
"Most praised" and "most criticized" labels shown on comparison and entity pages reflect the highest-scoring and lowest-scoring features in that phone’s Community Signals — not a separate judgment.
How We Write Community Insights
Community Insights summarize the strongest opinions that appear independently across multiple communities — not a single popular post, but a pattern repeated by different people in different places.
We look for shared meaning, not shared wording. If Reddit users say a battery "lasts all day," Quora users say they "rarely need to charge before bed," and XDA users describe "excellent endurance for daily use," we treat this as one recurring theme about battery life — even though the wording differs in each community.
Where communities disagree with each other on a topic, we do not manufacture a shared conclusion. A topic is only presented as a Community Insight when it reflects genuine agreement across sources.
All Community Insights are written in our own original wording. We do not reproduce, quote, or closely paraphrase individual posts or comments from any community.
The Community Summary Box and Discussion Volume
Every comparison, entity, and topic page includes a Community Summary Box showing the sources analyzed and, where applicable, the number of discussions behind that page’s data.
We only display a specific discussion count when it exceeds 1,300 analyzed discussions for that page. Below that threshold, no number is shown at all — we do not display an estimated or rounded figure. This is intended to ensure that any number shown reflects a substantial sample, rather than implying more data exists than we actually analyzed.
The Role of AI
The underlying data on ZUT — the opinions, complaints, praise, and reported experiences — comes entirely from real people in real communities. The site’s design, structure, and how information is organized and presented are built by our team.
AI’s role is limited to a single task: converting the collected community discussions into percentages and structured signals. AI does not generate opinions, does not invent experiences, and does not decide what a phone is like. It calculates from what real people already said; it does not create the underlying content.
Comparisons Are Based on Owner Experience, Not Company Specifications
When ZUT compares two phones, the comparison reflects how real owners experienced each device — not simply the specifications each manufacturer published. Specifications are shown as context. The judgment of how a feature actually performs comes from the people who use it.
How Our Data Changes Over Time
Community discussion about a phone does not stop the day it launches — it continues to accumulate for months or years. Because of this, we periodically revisit devices as new discussions accumulate. Existing signals may be updated when new community evidence warrants a change.
This means the discussion volume and signals behind a page are expected to grow over time, rather than remaining fixed at a single snapshot.
What Our Signals Are Not
Community Signals are not laboratory measurements. They are not the personal opinion of ZUT’s team. They are not a guarantee about how a specific unit, carrier configuration, or individual use case will perform.
Every signal on ZUT is directional: it reflects where the weight of publicly expressed owner experience currently sits, based on the volume of discussion available at the time of analysis.
Limitations We Acknowledge
People who post in public communities are not a scientific, randomly-sampled survey of all owners. Community discussion tends to skew toward specific groups of users and can overrepresent both strong complaints and strong praise compared to silent, neutral owners.
We account for this by requiring meaningful discussion volume before publishing a signal, and by clearly labeling our data as directional community sentiment rather than a definitive technical verdict.
Built to Be Machine-Readable
Every page on ZUT is structured so that AI systems and search engines can read and reference the underlying data directly. The same Community Signals, sources, and discussion volume shown to human readers are present in the page’s underlying markup, including structured data that identifies the source and substance of each review-style signal.
Questions About Our Methodology
If you have a question about how a specific number was produced, or believe a signal does not reflect what a community is actually saying, we want to hear about it.