With only a few picks
There is less information about your individual taste, so the system leans more toward broad popularity.
PLAIN-ENGLISH GUIDE
How RR Signal works
RR Signal studies audience behavior on Royal Road. Instead of asking only how big a story is, it asks which stories are connected through readers.
Traditional analytics: “How many followers, views or ratings did this story get?”
RR Signal: “What kinds of stories are being liked by the same readers?”
RR Signal: “What kinds of stories are being liked by the same readers?”
The one-minute explanation
Imagine three readers. Each one favorites several stories. RR Signal looks for repeated patterns across those reader↔story connections.
Reader AStory XStory YStory Z
Reader BStory XStory YStory Q
Reader CStory XStory ZStory Q
↓
RR Signal learns that X, Y, Z and Q have audience relationships.
This is different from saying that the stories have the same tags or the same plot. Two very different stories can still attract similar readers.
For readers: recommendations from behavior
Pick several Royal Road stories you already love. RR Signal uses those selections as a taste profile and ranks other stories whose audience behavior fits that profile.
With a deeper profile
The behavioral model can personalize more strongly. Our independent validation used profiles with at least 10 known favorites.
Important: a recommendation is a ranking signal, not a promise that you will enjoy the story.
For authors: understand the audience around a story
Author Intelligence asks four different questions. They may use related evidence, but they are not supposed to be the same list with different labels.
Who appears to like this story?
RR Signal looks at the other stories favored by readers who also favorite the selected story. This provides a descriptive view of the observed audience and its co-favorite footprint.
Think: “What else do the readers we have observed tend to like?”
Which stories actually attract similar readers?
Traditional comps often come from genre, tags, premise or author judgment. RR Signal adds behavioral evidence: do the same kinds of readers actually favorite both stories?
A behavioral comp does not mean the stories have identical plots. It means their observed audiences appear related.
Which nearby audiences are you not reaching yet?
A story can sit close to your audience while still having many observed readers who have not been seen on your story. RR Signal treats these as possible adjacent audiences worth investigating.
An Audience Gap is an opportunity hypothesis, not a guarantee of conversion.
Who may have compatible but non-identical reach?
The closest comp is not always the best collaboration opportunity. If nearly every observed reader already overlaps, there may be little new audience to reach.
Collaboration ranking considers audience compatibility, observed reach and possible non-overlapping audience.
Why a comp and a collaboration target are different
TRUE COMP
High similarity
Your story● ● ● ● ● ● ● ●
Story A● ● ● ● ● ● ● ○
Very similar observed audiences.
COLLABORATION OPPORTUNITY
Compatibility + new reach
Your story● ● ● ● ○ ○ ○ ○
Story B● ● ○ ○ ● ● ● ●
Some bridge, plus more observed audience outside the overlap.
Why some results say “early hypothesis”
Two observed readers are not the same as fifty. RR Signal therefore limits the strength of its language and, for very sparse stories, withholds collaboration recommendations entirely.
| Observed audience support | How RR Signal treats it |
|---|---|
| 0–2 readers | Early observations only |
| 3–4 | Early signals |
| 5–9 | Exploratory intelligence |
| 10–24 | Useful behavioral evidence |
| 25–49 | Good evidence |
| 50+ | Stronger evidence within the observed graph |
These are product guardrails, not universal statistical laws. Confidence should improve as the underlying
audience graph grows.
How this differs from a normal stats tracker
Story → metrics
Followers, views, ratings, rank, velocity and other performance measures.
Readers ↔ Stories ↔ Readers ↔ Stories
Audience relationships, behavioral similarity, adjacent audiences and potential audience bridges.
Both layers are useful. RR Signal is deliberately focused on audience structure rather than trying to replace every existing Royal Road analytics tool.
What RR Signal does not claim
- A Favorite does not prove that a reader finished or loved every chapter.
- A behavioral comp does not prove two stories are similar in plot, prose style or quality.
- An Audience Gap does not mean those readers will definitely enjoy your story.
- A Collaboration Target does not mean another author will want to collaborate.
- Results describe the audience data RR Signal has observed, not every Royal Road reader.
- Author Intelligence is still experimental and has not been validated as strongly as the reader recommendation endpoint.
Data & privacy
Behavioral research
The research graph uses publicly observable reader↔story relationships such as public Favorites. The behavioral model does not need chapter text to learn those relationships.
Your use of the site
Normal product feedback uses a pseudonymous browser identifier. Optional taste-profile contributions and private author-beta contact details require separate, explicit opt-in.
Author-beta email addresses are stored separately from the behavioral research database. The beta signup can be removed from the browser that created it.
FAQ
Does RR Signal read my chapters?
The behavioral model is based on reader↔story relationships, not chapter-text analysis.
Does a behavioral comp mean the stories are similar?
Not necessarily in plot or writing style. It means the observed audiences appear behaviorally related.
Why does my story have low evidence?
RR Signal has not yet observed enough qualifying readers connected to that story to make stronger claims.
Why isn't a story available?
The current live model contains only stories represented in the frozen research catalog used by the product.
Why do I need several stories for reader recommendations?
More selections provide a clearer taste profile. The strongest independent validation used at least 10 known favorites.
What is an Audience Gap?
A behaviorally adjacent audience that appears less represented among the readers currently observed on your story.
How is a collaboration target different from a comp?
A comp prioritizes audience similarity. A collaboration target also values possible non-overlapping audience reach.
Is RR Signal affiliated with Royal Road?
No. RR Signal is an independent research/product project and is not affiliated with Royal Road.
Want the technical version?
The methodology page explains the model, validation design, evidence status and the safeguards used to keep recruitment clues out of benchmark evidence.