Travel has been changed by comparison and review sites. Now it is fighting back by trying to harness big data to its transactional information in order to refine its proposition. David Reed spoke to Pascal Clement of Amadeus about how the sector is coping
You are going on holiday with your family and have decided to visit Capri. You think of yourself as an independent traveller and have a lot of knowledge of technology and social networks. So you decide to book the whole thing yourself – flights, hotel, car. Then just before your set off, it occurs to you to check on the details of ferries to get your hire car across to the island. That’s when you find out that no cars are allowed on Capri.
A travel agent would have known that, of course, but a travel site structured to find the best deal for each part would not identify the flaw in your plans. That real-life example comes from Pascal Clement, head of travel intelligence at Amadeus, who found out just in time that his best-laid travel arrangements would not quite work out the way he expected.
It illustrates the challenges facing the travel industry. Price comparison, deal and review sites have dramatically altered the way travellers think about their purchases. Consideration now starts with a look at who has the cheapest room or flight on offer, cross-checked against what others have said about that destination or airline.
“Social is massive and allows you to see how much people are choosing travel based on the comments they see,” says Clement. This has both upsides and downsides for travel companies. Positive comments and reviews provide the credibility of peer review to whatever the business says about itself through marketing. On the downside, “anybody can comment without having visited a hotel. I’ve had my own experience of choosing a hotel based on the comments made about it and finding it was the wrong decision – it was not what the comments were saying it was,” he says.
This disconnect between experience of a product and sharing views about it is one of the big data opportunities facing the travel industry. If companies were able 
to match commentators and reviewers to their customer base, it would make it easier to decide how to respond.
It has already triggered one entirely new business model in Hipmunk. Founded by one of the originators of Reddit, the travel website uses proprietary algorithms that consider a range of criteria to generate rankings, rather than just price. For airlines, it has created an Agony Index which considers the length of a flight and number of stops, while for hotels, its Ecstacy Index brings together price, facilities and reviews.
Amadeus is working on its own quality and satisfaction index based on customer experiences. “It is an area where we can bring in the basics, such as how long it takes to get to a destination, the distance of a hotel from the airport to create a rating,” says Clement.
QSI is one of the initiatives that he was hired to support as the company works on developing new data sources and technology platforms for its clients in the sector. Another is the report it sponsored, “At the Big Data Crossroads: turning towards a smarter travel experience,” which was written by analytics in business guru Thomas H. Davenport based on interviews with 21 key players in the industry.
Clement points out that while big data and social network analytics get a lot of attention, “there are so many things to improve at the basic level before companies even think about a revolution in their approach. A lot of the basics have never been put in place in the travel industry – which is a great opportunity for us to help customers.”
In some areas, travel companies are at the leading edge of the use of analytics. Pricing and revenue management are examples where airlines and hotels bring very powerful skills to bear on deciding how to balance the need to fill as much inventory as possible with optimising revenues and also protect future price points. There has been something of a cold war between providers and comparison sites in this sphere which is creating a need for more innovation.
“There is no big difference in the nature of the big data available in the travel industry – operational, business, social, local and mobile. But there are some additional things that are interesting, such as sensors on board airplanes. Everything the travel industry touches produces data,” he says.
To date, most of the usage made of this data has been relatively traditional, rather than transformational. There are initiatives underway to move things on, such as BA’s “Know Me” programme wh
ich is looking to integrate existing data from the Executive Club loyalty scheme with big data from social networks. The goal is to recognise customers better, enhance service and recover from any failures, and to develop more inspiring offers. Listening to what is being said in the social space and harnessing it to a proven CRM platform ought to be highly effective.
Clement admits that this is a rare example. “The report found that because of the complexity, cost and time issues, there are very few places where we could see a strategy of putting all that data together. Most sources are very siloed, even if they are very complementary,” he says.
As with most other sectors now trying to leverage extended data assets, there is a skills issue. “I’m not sure the travel industry has any more of an advantage in hiring in data scientists which are in shortage all over the place. That is a real problem,” he says.
Among the recommendations made by Davenport in his report are to start assembling big data skills, either through internal hiring or by working in partnership with vendors. He notes that it is likely travel companies will have to work with partners in order to realise their big data goals, given that most companies are only delivering part of the experience, even though the customer wants their journey or holiday to be seamless.
Being one of those vendors and partners is an opportunity which Amadeus is exploring, including through the creation of a new technology platform to offer data integration. This is one area in which clients struggle as there are often not the skills in merging structured and unstructured data into a single database.
Another challenge is whether genuine data scientists will be able to deliver meaningful insights to an industry with very specific requirements. “There is a big issue around combining the knowledge of a pure data scientist with deep knowledge of travel. We are in the process of trying to understand if it is better to bring a data scientist on board and teach them the travel industry or to take people with a deep knowledge of travel and teach them data science,” he says.
Like many other sectors, the demands and behaviour of the customer have run ahead of the capability of providers who are now trying to join up those dots. Given the appetite for information, there is clearly a competitive advantage in finding ways to join social and transactional data to drive better propositions.
For Clement, there is no time to lose. “Travel is late into analytics and its skills are not strong. What we need to bring is people with a deep knowledge in analytics and use our knowledge of travel to help clients get there,” he says.
