Listings scored
23,233
public + derived
TA-P1-001 · after cleaning
Project 1 · Content operations framework
Content Quality Audit
Decision: which incomplete Bangkok listings to fix first under limited ops capacity. Stakeholder: Product Operations / content ops. Equal-weight portfolio case — not the flagship.
Overview
Below score 70
11,480 (49.41%)
derived
TA-P1-002 · provisional threshold
Top-50 review uplift
8,908
modelled review proxy
TA-P1-003 · not observed bookings
Critical tier
2,701
derived
Score < 40
Diagnosis
Description gap
94.46%
derived
Below 150 words
Reviews gap
53.03%
derived
Below 5 reviews
High vs low median reviews
16 vs 0
derived association
Association, not causation
Regression association
~77% / +10 pts
derived association
R² = 0.4515 · holds price
Working-assumption weights (sum 100): description 35, amenity 25, host 20, reviews 20. Host response missing: 2,742 ( 11.8%). Cleaning retained 23,233 of 28,806 source rows.
Priority and action evidence
| Segment | Listings (n) | Avg score | Top gap | Recommended action |
|---|---|---|---|---|
| Premium High Potential | 2,870 | 49.62 | description | Immediate 1:1 outreach. Assign content specialist. Upgrade photos/description within 7 days. Highest ROI per effort. |
| Mid Market Volume | 6,296 | 52.04 | description | Batch content templates by gap type. Weekly sprint of 20–30 listings. Use standardized amenity/description checklists. |
| Long tail Low Priority | 2,314 | 47.87 | reviews | Monitor quarterly. Auto-email hosts with self-serve content guide. Do not allocate manual ops until score drops to Critical. |
Recommended action: prioritize Premium High Potential outreach, then Mid-Market description batch fixes. Top-50 modelled impact: 8,908 modelled review proxy (not an observed outcome).
Weight sensitivity (working assumptions)
| Scenario | Weights | Below threshold | Tier-change rate | Spearman vs baseline |
|---|---|---|---|---|
| baseline_current | d35/a25/h20/r20 | 11,480 (49.41%) | 0.0% | 1.0 |
| equal_components | d25/a25/h25/r25 | 10,342 (44.51%) | 8.66% | 0.9862 |
| description_heavy | d45/a20/h20/r15 | 12,884 (55.46%) | 8.88% | 0.9755 |
| trust_heavy | d30/a20/h25/r25 | 10,869 (46.78%) | 7.72% | 0.9921 |
Segment stability tables with sample sizes live in
output/sensitivity_analysis.md.
Caveats
- Public Airbnb Bangkok snapshot (2025-09-26) — OTA industry analogue, not Agoda inventory.
- Score weights are working assumptions; sensitivity explores alternatives that sum to 100.
- Association between content score and reviews is not causal impact.
- The value 8,908 is a modelled review proxy, never an observed booking or revenue outcome.
- Offline artifact: no CDN scripts or stylesheets.