How the grading works
Every report grades one listing against a fixed, weighted rubric — the same rubric for every listing, so a B means the same thing everywhere. Two kinds of criteria feed it: factors Airbnb has publicly documented as ranking inputs (quality, popularity, price, and location), and conversion mechanics — what makes a guest who sees the listing actually book it.
The six dimensions
Visual Impact — 25% of the grade
Cover photo strength at thumbnail size, whether every guest-used space appears, running order, lighting, duplication, and whether captions sell benefits or merely label rooms. Photos are the first thing every guest judges, which is why this is the heaviest dimension.
Conversion Copy — 20% of the grade
Whether the title leads with a real differentiator, whether the description opens with a hook rather than a limitation, structure and scannability, and whether the obvious guest objections are answered before they're asked.
Discoverability — 20% of the grade
Coverage of filterable amenities — every undeclared amenity silently excludes the listing from that filtered search — plus completeness signals Airbnb's search rewards. Includes an audit of high-value amenities the listing may have but hasn't declared.
Completeness & Coverage — 15% of the grade
Trust & Social Proof — 12% of the grade
Rating and all six subscores, review count, badges, host response record, and the listing's most recent guest reviews — quoted verbatim, with negative signals weighted by recency: a complaint that stopped appearing was likely fixed; a recurring one is a live problem.
Friction & Accuracy — 8% of the grade
House-rule load, check-in flexibility, and signals that the listing does or doesn't match reality — accuracy gaps surface in reviews and in Airbnb's accuracy subscore.
What we actually look at
The full public listing: every photo in the gallery (not a sample), the complete title and description, the full amenity list, live pricing for sample dates including the fee breakdown, ratings and all subscores, badges, host response record, house rules, and the twelve most recent guest reviews in full.
What we refuse to do
- No guessed facts. Every claim in a report is verified against the captured listing. If a field couldn’t be read, the dimension is marked unscored — never estimated.
- No invented amenities in recommended copy. When a high-value amenity is missing from the listing, the report asks (“if the property has a dishwasher, declare it”) — it never writes unverified claims into text a host might paste.
- No fabricated dollar figures. Revenue impact appears only when the listing’s real pricing supports the math.
- No grading of bad data. Every capture verifies its own completeness against the listing’s claimed counts before any grading happens; incomplete captures are rejected, not graded.
- Verbatim review quotes only. Guest language is quoted exactly as written, and any stated count is arithmetically true of the reviews captured.
The grade scale
The scale is calibrated against real listings and re-checked whenever the underlying model changes. Grades judge the listing’s merchandising, never the property or the host — the most common finding across all reports is a genuinely excellent property undersold by its listing.