Expedia has launched its Expedia’s Innovation Lab for Asia located in Singapore, signalling its focus on transforming travel through innovation and application of technology. The transformation of travel for Expedia is not just about making the average experience better with every single click of the person before, but through analysis and data insights, also improving each person’s unique experience. By understanding the data-driven insights, this can be done quickly and seamlessly.

“Expedia’s global success comes from a culture that values innovation and investment in technology. The key differentiator is that we do every single improvement based on powerful data insights. It has been our driving force for the past 20 years.

“We invested US$1.2 billion in technology in 2016 and with our 2,000 engineers and 700 data scientists we deliver personalised experiences that come with more choice, flexibility and savings for both our customers and partners”, said Expedia group Product and Design senior VP Arthur Chapin (pic, above).

“We are already preparing for the time when machines will be able to not just analyse but also understand when to act on real-time data. The companies that are going to win in the end are those that structure themselves to prepare for these changes, act on them and be first to market”, he added.

Expedia sees a big opportunity in mobile. The company saw 180 million downloads of its mobile app in 2016 with nearly one third of all transaction and more than 65% of hotel room night bookings coming through mobile platforms in 2016.

The other areas of interest to the company are Messaging, Natural Language Processing and Machine Learning but it is the confluence of all these that Expedia believes will be where the magic happens.

“People don’t travel to go to a place, people travel for an outcome and the next big thing in travel will be a better use of information to make the travel inspiration and booking experience more personalised and predictive in real-time”, enthused Chapin.

Expedia places great emphasis on addressing the question on ‘what have they solved for their customers’ which is very much done around a hypothesis and test world – referred to ‘test and Learn’ internally.

This has changed the way the company thinks about innovation, having moved from four release cycles per year to weekly releases allowing for an agile development.

The company performs more than 1,500 tests a year and it believes that even if half of them fail, that still equates to hundreds of wins. This culture has been credited with company’s increased innovation.

Over the years Expedia has made a significant investment in data science in areas of selection and personalisation.  Aside from employing 700 data scientists, the company deploys selection algorithms as a “learning to rank problem”.

For example, the travel graph tracks the billions of searches done on the Expedia websites and the hotels that were clicked and purchased.

This is then combined with all kinds of information about the hotels – from pricing for that specific stay to popularity, star rating, location, segment preference and more.

Depending on the feature in questions, a number of cutting edge algorithms are used – from simple regression models to complex ensemble models using “boosting” methodology and various clustering schemes.

The traditional web interface for travel relies on customers coming to Expedia with some of their questions already answered. It’s a format that hasn’t fundamentally evolved in years.

Chatbots are potentially the next evolution in discovery, able to answer the real questions travellers have, such “where should I go,” or “what hotel is best for kids”.

With a chatbot interface currently available in the US and the UK, Expedia and the travellers have an opportunity to interact in a more open-ended and natural manner.

Rather than customers having to input line-by-line the details about a trip, chatbots allow for customers to ask and then answer questions in more open, conversation-like method. This helps alleviate the risks of over- or under-personalising content.

DNA