Three Ways TripAdvisor Uses Big Data To Become The World’s Largest Travel Site
The website TripAdvisor was founded in 2000 and is a travel website that offers reviews of travel-related content such as hotels, restaurants and attractions. The TripAdvisor Media Group, the owner of TripAdvisor, currently operates 25 travel brands, operating in 45 countries. The website is all about user-generated content and is free to use for its users. It is supported by and advertising business model, which means that more visitors will mean more revenue. Hotels can purchase a display link to their own website and the amount of click-throughs are what determines whether or not a subscription is renewed. More visitors means more click-throughs, means more revenue.
TripAdvisor therefore turned to Big Data to offer the best recommendations and content for its users. Currently they have over 280 million unique monthly visitors, over 170 million reviews covering over 4 million accommodations, attractions and restaurants from over 140.000 locations around the globe. It may be clear that TripAdvisor generates massive amounts of data, which they use to improve their service and to become the world's largest travel website.
Ensure Subscription Renewals
According to Wikibon, TripAdvisor has built advanced Big Data models to predict the amount of click-throughs that are required for a hotel to renew their subscription. They calculate the amount of marketing activity required to achieve a certain level of click-throughs, which would result in a subscription renewal. However, according to Director of Analytics Michael Barry, “for some clients […]even a significant increase in click-through rates only marginally increases its likelihood of renewing”. Using advanced analytics they know whether or not it is economically wise to increase the marketing activities.
The objective of TripAdvisor is to help users find better content more quickly. They achieve this by analysing their individual behaviour on the website as well as the activity on the visitor population as a whole. By offering a better experience, users will be more successful in finding what they are looking for, resulting of course in more advertising revenue in the end. They are using several Big Data techniques to achieve this, ranging from large-scale real-time analytics, predictive analytics, data mining and statistical modelling. All focussed on delivering better personalized recommendations for the visitor.
Personalized, Mobile, Content
But they are not only focussing on providing personalized content via the desktop. They are actively using Big Data analytics to improve mobile travel. Over 128 million travellers have downloaded the TripAdvisor App and they are working hard to improve the mobile experience as well. Their tablet monetization is similar to that of desktop visitors, but their smartphone monetization is just at 20% of desktop monetization. To improve this number, they are working hard to improve their metadata analytics capabilities. However, for this to work they need to increase their partners that allow direct booking via the TripAdvisor App, which is still rather limited as larger players have declined to work with them.
To offer a better mobile experience, TripAdvisor has developed augmented reality features. This is possible because they have massive amounts of data thanks to the reviews, but also factual points of reference about a city, community data and the tourist-related services being used by travellers can be incorporated. If they are capable of combining data from previous visits on other channels, with all this valuable data, they can start offering truly personal recommendations based on the profile and location of the user. However, at the moment TripAdvisor is not this far yet, although according to Joost Schreve, vice president of mobile at TripAdvisor, they “… will get there”.
Combat Fraudulent Reviews
Another Big Data application used by TripAdvisor is to beat fraudulent reviews. Fake reviews are harmful for entrepreneurs, are useless to visitors and in the long run will negatively impact TripAdvisor. One solution to beat fake reviews is to work with verified reviews, where it is confirmed, via a 3<sup>rd</sup> party integration, that a reviewer actually stayed at a certain property. Beginning of 2014, TripAdvisor and American Express sealed the deal that enables Amex cardholders to connect their card to their TripAdvisor profile. Whenever they leave a review on the website, it is market as an “Amex Card Member Review” confirming that the card was used to make a purchase at those locations. Such verified reviews are only one part of the deal. According to a spokesperson of TripAdvisor, they have “extensive algorithmic protection in place to prevent fraud”.
A Custom Big Data Platform
In order to deal with all their data at hand, to develop personalized recommendation and improve the anti-fraud algorithms, TripAdvisor has developed a custom Big Data platform. They use a myriad of Big Data technologies and techniques including Hadoop (to store and process web log data), SQL Servers (to report against aggregated data), Hive (to query the data and put it into tables), Machine-learning to continuously improve the site experience as well as other technologies such as Redshift, R and Python.
TripAdvisor sits on a massive pile of data from all their visitors and the reviews on their website. Combined in the right context, this can result in a great experience for the visitors and additional revenue for TripAdvisor’s business partners. They are continuously working on improving their Big Data platform and are hiring extensively to achieve that. For TripAdvisor, as well as other travel giants, Big Data is the only way forward to continuously offer visitors a great experience and in the case of TripAdvisor, to become the world's largest travel website.