Data‑Driven Product Design at Ctrip: Practices and Case Studies

Data and design have long been debated in the industry, with terms like “data‑driven design” and “data‑guided design” circulating alongside concerns that data focus may weaken user experience. This article explores how Ctrip applies data throughout its product design lifecycle, using two concrete cases to illustrate practice.

Product Design’s Data View – Designers must adopt a correct “data view”: data permeates the entire design process, is created by people and interpreted by people, should be collected and analyzed based on clear business or design goals, monitored after launch for rapid iteration, continuously evaluated, and never replace user experience.

Effective data analysis requires solid data sources and a strong “data sense” – the ability to detect anomalies and derive insights, as exemplified by historical anecdotes of intuitive data interpretation.

At Ctrip, data analysis is a core competency, influenced by founder Liang Jianzhang’s computer background and Oracle experience. Data informs every stage: from diagnostic research in the ideation phase, value estimation and goal setting in design, massive A/B testing and monitoring at launch, to ongoing operational dashboards.

Ctrip has built several data tools:

UIP User Insight Platform – aggregates key metrics (PV, UV, bounce rate, conversion, dwell time) across channels, pages, locations, and traffic sources.

Page Click Plugin – provides click summaries, trends, detailed statistics, browser breakdowns, and heatmaps.

A/B Testing – scientifically samples traffic to compare design variants, yielding conversion‑rate insights; widely used in the hotel product line with success rates above 15%.

KPI Portal – custom dashboards that integrate project‑specific metrics and even translate them into revenue estimates.

When quantitative tools fall short—e.g., assessing user preferences for room‑type information on hotel detail pages—Ctrip’s user research team steps in, employing interviews, focus groups, card sorting, and eye‑tracking to uncover qualitative insights.

Case 1: Evolution of Ctrip’s Homestay Channel

The homestay channel was launched to capture the growing market for boutique accommodations. Initial designs mirrored standard hotel templates, resulting in lower conversion rates. Data revealed that homestay users booked earlier and traveled farther, indicating a vacation‑oriented mindset.

Based on data mining, user interviews, and competitive analysis, the redesign focused on:

Home page – emphasizing browsing and recommendations, reducing search prominence, using warm visual tones.

List page – offering large‑image mode to showcase homestay characteristics.

Detail page – exposing partial content of each module for preview.

First‑iteration A/B tests showed mixed results: overall order conversion improved, but home‑to‑list conversion dropped 20% (search visibility issue), list‑page conversion rose 10% (large images underperformed), and detail‑to‑booking conversion increased (module exposure helped decision‑making).

Two weeks later, a second iteration reversed the home‑to‑list decline (+2%) and lifted overall order conversion by 26%.

Case 2: Kezutong App Order‑Detail Page Optimization

Kezutong, a B‑side app for homestay owners, originally displayed order details line‑by‑line, forcing users to scroll through three screens. Monitoring showed only 13% of users edited orders, with most edits focused on order status, and an average dwell time of 11 seconds per page.

The redesign reorganized information into four modules (booking, guest, room, settlement), streamlined high‑frequency status edits, and hid low‑frequency content edits behind an entry point. Usability testing involved high‑fidelity prototypes and on‑site observation of owners.

Post‑launch metrics showed over 30% increase in page views, average edit time reduced from 11 seconds to 4 seconds, and a 30‑fold rise in order‑edit volume, indicating that the streamlined mobile experience successfully shifted owners from PC to mobile.

Conclusion

Both cases demonstrate that continuous data monitoring and analysis, combined with qualitative user research, guide product iterations throughout the lifecycle. Effective data‑driven design starts with clear questions, targets specific users and processes, defines measurable outcomes, and acknowledges cross‑impact effects. Ultimately, data must serve human‑centered design, ensuring that insights translate into better user experiences and business results.