Former Cool News real estate team make outbound travel decision engine "word of mouth"

Week is easy to be in charge of real estate search business at Cool News first. After the cool news experienced turbulence, a comprehensive transformation of travel search, he left from Cool news, began the journey of entrepreneurship. At that time, he found that the tourism industry, although through the search engine to solve the problem of low prices, through the product part of the solution to the problem of content precipitation, but the real rise to the level of consumer decision-making, but also lack of substantive content. So he began to do the tourism community "517 Travel", want to pass the UGC way to the consumer to precipitate some subjective, has the practical operation value the reference suggestion.

In general, however, in the year of the PC end of the tourism content precipitation, 517 travel did not have to do the ma Honeycomb, poor travel, and in the later mobile end, 517 travel and slow travels of the travel category application bread, on the road. So Zhouqingsong began to think about the transformation of the team. Starting in January this year, Zhouqingsong decided to let the team pick up the old line of search and try to do it in a different way.

In Zhouqingsong's view, social platforms such as Facebook and Twitter have long been precipitated by huge data mines, and the time has come to make travel comments using search engines (UGC). And alone in the vertical field of tourism, TripAdvisor sedimentation a lot about the hotel's reliable evaluation, yelp in the restaurant evaluation to do the authority, Priceline in the field of car rental is also doing a good job.

As a decision engine, Word-of-mouth Travel has three main tasks. The first is data crawling. Word of mouth travel can basically do the whole network travel review data crawl, 70% of the data source from the foreign community (this is mainly due to the first step of Word-of-mouth travel to locate the largest piece of tourism cake-outbound travel). The second step is data structure. It contains two dimensions of transverse longitudinal. Horizontally, the engine will classify the data in vertical categories, including restaurants, hotels, car rentals, shopping, activities, etc. Vertically, it distinguishes the information from the Evaluation class information (including the appraisal level information) and the descriptive information (travel experience Class), then give the different information to pay the corresponding weight, the end of the specific travel products (such as a restaurant) comprehensive scoring, at the same time, "praise", "evaluation", "Difference", "unclassified" to the information presented to the user. Finally, Word-of-mouth Travel has not yet released the function-personalization. In short, the next edition of Word of mouth travel will try to understand your travel preferences in addition to analyzing the reputation of travel products in the user community. In other words, the more you use, the more your recommendation will suit your taste.

From the start of this year January project, to March Word-of-mouth travel Thai version online, the team took only 3 months. Next, the team will launch the American and French editions.