How Recommendation DNA Works
Recommendation DNA is an analysis framework utilized for scoring audiobooks based on a set of quality dimension criteria (which will be explained below). Additionally, we use what we call “listening-profile” and “listener-fit” indicators to provide and explain what the audiobook experience may feel like and who might be best suited for that audiobook. These set of criteria are applied the same across all the audiobooks.
What Recommendation DNA Measures
Recommendation DNA separates quality, experience, and fit so readers can understand what the score means at a glance.
Quality Dimensions
These determine the Recommendation Score itself.
- Eight weighted quality dimensions
- Narration and story carry the most influence
- Same framework for every audiobook
Listening Profile
These describe what the listening experience feels like.
- Pacing, tone, humor, romance, complexity, and more
- Higher score = more prominent, not better or worse
- Helps readers understand the experience
Listener Fit
These help show who the audiobook may suit best.
- Commute, road trips, bedtime, workouts, and more
- Shows ideal listening situations
- Does not affect the quality score
Book Evidence vs. Audiobook Evidence
Recommendation DNA separates evidence about the underlying book from evidence about the specific audiobook recording being evaluated.
Book-Level Evidence
Book-level sources can support dimensions such as story quality, character development, originality, world building, and other narrative traits.
Audiobook-Specific Evidence
Narration and audio-production claims require evidence tied to the correct narrator and recording, or to a verified equivalent audiobook edition.
What Contributes to the Recommendation Score
These eight quality dimensions make up 100% of the score.
Example: How a Score Is Calculated
Sample audiobook with a final score of 8.7/10.
| Quality Dimension | Rating | Weight | Contribution |
|---|---|---|---|
| Narration Quality | 9.4 | 17.8% | 1.67 |
| Story Quality | 8.8 | 17.1% | 1.50 |
| Character Development | 8.3 | 12.5% | 1.04 |
| Emotional Impact | 8.7 | 11.8% | 1.03 |
| Audio Production | 9.0 | 11.2% | 1.01 |
| Originality | 8.1 | 10.5% | 0.85 |
| Replay Value | 8.0 | 9.9% | 0.79 |
| World Building | 8.5 | 9.2% | 0.78 |
| Total | — | 100% | 8.67 → 8.7 |
When a Score Is Withheld
Recommendation DNA does not force a numerical rating when the available evidence is too limited, too concentrated in one source network, materially conflicting, or insufficiently tied to the correct audiobook recording. In those cases, we may publish a developing listening profile while leaving the overall Recommendation Score unrated. “Not yet rated” does not mean zero or poor quality; it means the evidence is not yet strong enough to support a responsible numerical result.
Recommendation Score: What It Means
Evidence Confidence: How Strongly the Score Is Supported
Evidence Confidence does not change the Recommendation Score. It describes how strongly the result is supported based on source independence, recording and edition matching, evidence coverage, consistency, and meaningful disagreement.
High Confidence
75–100%
Broad, consistent support across multiple independent sources.
Moderate Confidence
55–74%
Useful supporting evidence is available, but important gaps remain.
Limited Confidence
Below 55%
The result has meaningful evidence gaps or limited independent support. This does not mean the audiobook is low quality; it means the current evidence base is not yet strong enough for high confidence.
Research & Open Data
The Audiobook Listener-Fit Framework is also available through public research and data repositories for transparency, citation, and reuse.
View the Recommended Audiobooks research hub
Explore the public listener-fit dataset
What Readers Should Remember
Shows how strongly the audiobook performs across the quality framework when the evidence is sufficient to support a numerical result.
Shows what the listening experience feels like and can still be published as a developing profile when a numerical score is withheld.
Shows how firmly the result is supported by independent evidence.
Help readers understand not just whether a book is good, but who it may be right for.