AO3 is the largest fanfiction archive on the internet, and it has no rating system. No stars, no scores, no "best of" ranking. This is deliberate — the Archive's founders wanted to host works, not rank them. But readers still need to find good stories among 12 million works. So they've developed proxy metrics using the data AO3 does provide: hits, kudos, comments, and bookmarks.
Each metric tells you something different. None of them are reliable quality indicators on their own. Understanding what they actually measure is the key to effective AO3 discovery.
What each metric means
Hits
The number of times a work has been viewed. A hit is counted every time someone loads the work page, including repeat visits from the same reader across multiple chapters.
What it measures: Exposure and traffic. A work with many hits has been seen by many people. This could mean it's popular. It could mean it's been on AO3 for ten years. It could mean it's controversial and people keep clicking to see what the fuss is about. It could mean it has a premise that attracts clicks but doesn't necessarily retain readers.
What it doesn't measure: Quality. A work with 100,000 hits and 50 kudos has been widely seen and not widely appreciated. A work with 5,000 hits and 800 kudos has a small audience that loves it.
When to use it: Almost never as a primary quality signal. Hits are too correlated with age and chapter count to be useful for comparison.
Kudos
AO3's equivalent of a "like" button. A reader clicks the heart icon to leave kudos on a work. A reader can only leave kudos once per work, regardless of how many chapters they read.
What it measures: Reader appreciation. A work with many kudos has been enjoyed by many readers. Kudos are the closest AO3 has to a quality metric.
What it doesn't measure: Intensity of appreciation. Kudos is binary — you gave it or you didn't. It doesn't distinguish between "this was fine" and "this changed my life." The one-kudos-per-work limit also means multichapter works don't accumulate more kudos from repeat readers, which biases the metric toward one-shots. A reader who reads and enjoys a one-shot leaves one kudos. A reader who reads 50 chapters of a longfic and loves it also leaves one kudos. These two reading experiences are not the same, but kudos treats them identically.
When to use it: Sort by Kudos when you want the most broadly appreciated works in a fandom. It's the best general-purpose quality proxy. Expect older works and one-shots to be overrepresented.
Comments
The number of comment threads on a work. AO3 counts threads, not individual comments — a conversation with 20 replies counts as one comment thread.
What it measures: Engagement. A work with many comments has sparked discussion, emotional reactions, or both.
What it doesn't measure: Whether the comments are positive. A controversial work can have many comments, most of them negative. A work can have many comments because the author replies to every single one, creating long threads. The comments metric is the noisiest of AO3's engagement signals.
When to use it: As a secondary signal alongside kudos. A work with many kudos and many comments is broadly appreciated and deeply engaging. A work with many kudos and few comments is broadly appreciated but hasn't sparked discussion — which is fine, especially for porn or short fluff.
Bookmarks
The number of readers who have bookmarked a work. Bookmarks are public by default (readers can make them private).
What it measures: Intent to return. A reader bookmarks a work when they want to find it again — to reread, to recommend, to wait for updates. This is a stronger signal than kudos because it requires more intent.
What it doesn't measure: Bookmarks accumulate more slowly than kudos because fewer readers bookmark than leave kudos. A newer work will have disproportionately few bookmarks relative to its quality.
When to use it: Sort by Bookmarks when you want the works that readers consider worth returning to. This tends to surface long, well-written works and filter out one-shots and lower-effort pieces. For finding novel-length quality fiction, Bookmarks is often a better sort than Kudos.
Ratios that tell a story
The relationship between metrics is often more informative than any single number.
High kudos-to-hits ratio: A work with 10,000 hits and 1,500 kudos (15% ratio) is doing well. Readers who find it tend to like it. This is a common pattern for quality works in smaller fandoms or niche tags.
Low kudos-to-hits ratio: A work with 100,000 hits and 500 kudos (0.5%) is being seen but not appreciated. This can signal a work that has an appealing premise but poor execution, or a controversial work that people click on but don't like.
High comments-to-kudos ratio: A work with 500 kudos and 200 comment threads is generating significant discussion. This often signals a divisive or thought-provoking work, or an author who actively engages with every commenter.
High bookmarks-to-kudos ratio: A work with similar bookmarks and kudos counts has a dedicated audience that considers it reread-worthy or reference-worthy. This is common for long, plotty works — readers bookmark them to revisit specific scenes or keep track of complex plots.
The age bias problem
All AO3 metrics favor older works. A 10-year-old fic in an active fandom has had a decade to accumulate hits, kudos, and bookmarks. A 2-month-old fic by the same author at the same quality level will have drastically lower numbers.
Sorting by any metric biases results toward older works. The fandom's "all-time greats" dominate the top of every sorted list, and they stay there because visibility drives engagement, which drives visibility.
To find newer works: sort by Date Updated and look at the first few pages for works with unusual engagement relative to their age. A 2-week-old work with 200 kudos is outperforming a 10-year-old work with 2,000, but the raw numbers won't tell you that.
What AO3's metric system gets right
For all the limitations, AO3's approach has advantages over star ratings or review scores:
- No review bombing (no scores to manipulate)
- No rating inflation (every kudos is equal, no 1-5 scale)
- No gaming for visibility (metrics don't control what appears on browse pages)
- Deliberately imprecise (the Archive doesn't pretend to measure "quality," just engagement)
The system trusts readers to develop their own quality heuristics rather than outsourcing that judgment to a single number. That trust is earned — AO3 readers become more sophisticated evaluators because the platform doesn't do the evaluation for them.
If you're looking for a specific kind of quality — long, complete, well-liked, rereadable — the combination of filters (Complete, minimum word count) and sorts (Bookmarks) gets you closer than any star rating would.