Travel Product & Supply Analytics
Four equal travel-marketplace cases: content triage, funnel experiment, weekly ops packaging, and public-source readiness operations.
MY ROLE
Sole analyst across content scoring, funnel experimentation, weekly ops packaging, and the Bangkok content-operations platform case.
THE DECISION
Which content gaps, experiment readout, weekly exceptions, and readiness tickets should a Product Ops analyst act on first—under explicit evidence labels?
FOUR EQUAL CASES
Each case carries its own decision, evidence label, finding, action, and limitation.
Content Quality Audit
Decision. Which public-listing content gaps should be addressed first?
Evidence label. Public source + derived metrics; proxy opportunity estimate
Finding. 11,480 of 23,233 listings are below provisional score 70
Action. Prioritize the ranked backlog; validate outcome with an experiment
Limitation. The 8,908 annual-review estimate is a proxy, not booking, revenue, or causal impact
Conversion Funnel & Experiment
Decision. Should the PDP intervention ship?
Evidence label. Synthetic clickstream + simulated experiment
Finding. 25,000 sessions, 77,096 events; simulated result is DO NOT SHIP
Action. Keep the control and redesign before retesting
Limitation. No real user behavior or realized lift is represented
Weekly Product Operations
Decision. What exceptions need an owner in the next weekly cadence?
Evidence label. Derived portfolio package + simulated workflow
Finding. Reconciles P1 content and P2 funnel/experiment outputs
Action. Assign owners to the action queue and review exceptions weekly
Limitation. Illustrative WoW and workflow; no verified time-saving claim
Bangkok Content Operations Platform
Decision. Which public-source readiness issues should enter the ticket queue?
Evidence label. Public-source analogue + derived metrics + simulated workflow
Finding. 14,551 of 31,069 listings are below provisional readiness score 70
Action. Triage the simulated ticket queue by priority and SLA
Limitation. Airbnb/OSM analogue, provisional threshold, and simulated ticket/SLA fields; not Agoda data or workflow
APPROACH
- Scored public Bangkok listings and ranked a content backlog with demand proxies
- Built synthetic funnel SQL and a simulated PDP experiment readout
- Packaged P1/P2 outputs into a derived weekly ops demonstration
- Modeled public-source readiness issues with a simulated ticket/SLA layer
EVIDENCE & DELIVERABLES
- 11,480 of 23,233 listings below provisional content score 70 (derived)
- 25,000 sessions / 77,096 events; simulated experiment decision DO NOT SHIP
- Weekly ops package reconciles P1 content and P2 funnel/experiment outputs
- 14,551 of 31,069 listings below provisional readiness score 70 (derived)
SELECTED WORK FILES
Travel Analytics Case Library
Four equal cases with linked dashboards, executive briefs, methodology, limitations, and SQL where applicable.
8 KBTravel Analytics Portfolio Case
A 12-slide portfolio overview that connects the four cases without elevating any single case.
216 KBWeekly Ops Diagrams
Clickable diagram previews: open each full diagram or its Mermaid source file. Sample/process only — not Agoda production.
20 KBTravel Funnel Dashboard
Browser-ready conversion funnel and experiment dashboard built from synthetic clickstream data.
5 KBWeekly Product Ops Brief
Slack-ready weekly brief combining content health, funnel metrics, experiment readout, and ops flags.
4 KBContent Quality Prioritization SQL
DuckDB queries that size content gaps and rank public-listing improvement opportunities; no listing data is bundled.
4 KBFunnel & Experiment Analysis SQL
DuckDB funnel, device, cohort, and experiment-population queries for the synthetic clickstream case; no event data is bundled.
7 KBEVIDENCE INTEGRITY
This case is presented as public listings + synthetic clickstream · proxy and simulated impact. The label distinguishes observed work from simulated impact, proxies, historical comparisons, or proposed architecture.