Travel Analytics case library · offline portfolio
Each case answers a different operating question. Evidence labels distinguish public-source analysis, synthetic data, proxy estimates and simulated workflows.
Portfolio boundary: independent portfolio cases, not Agoda production data, revenue, bookings, workplace performance or causal business impact.
Evidence
public derived review proxy
Finding
11,480 / 23,233 listings (49.41%) below provisional score 70.
Action
Review the highest-priority RICE content gaps.
Limitation
8,908 is a modelled review proxy—not bookings, revenue or causal uplift.
Evidence
synthetic simulated experiment
Finding
DO NOT SHIP: 3.64% vs 3.66%; p=0.936120; CI −0.59pp to +0.64pp.
Action
Do not ship on this simulated primary readout.
Limitation
25,000 sessions and 77,096 events are synthetic; design remains underpowered.
Evidence
mixed upstream simulated workflow
Finding
P1 content and P2 funnel/experiment artifacts reconcile in one weekly-style package.
Action
Keep content triage moving and retain P2’s simulated DO NOT SHIP decision.
Limitation
Illustrative WoW is not live; time-saving claim is unavailable and omitted.
Evidence
public derived simulated workflow
Finding
14,551 / 31,069 listings (46.8%) below provisional readiness 70.
Action
Run a two-lane remediation pilot; retain entity matches for human review.
Limitation
Tickets/SLA are deterministic simulations; reviews are a demand proxy, not bookings or revenue.
Portfolio case
Travel Analytics case overview (PDF) · 12-slide portfolio overview; it does not elevate any one of the four cases.