It’s often said that it is difficult to make predictions, especially about the future. Statistics help a lot with this, such as estimating the volume of people likely to be travelling from Abaeté to Zyryanka during March, or the number of hot breakfasts that will be ordered on an early morning flight from Johannesburg to Doha. Predictive analytics using statistical models have been used to maximise yield and predict capacity in the travel industry for many years.
However, it is more difficult to predict the likelihood that a given traveller will purchase that breakfast, need that seat with the extra room or which hotel they need to stay in. Yet this is what modern travellers expect. Their experience with online retailers is personalised to their wants and needs and delivered at the right time and in the right way.
A major trend in predictive data analytics over the past few years has been the adoption of machine learning to reduce this difficulty. Machine Learning is a type of Artificial Intelligence (or AI) that relies on detecting patterns in large amounts of data. AI has had a rocky past, with much promised by researchers but not a lot delivered. Lately though, AI has started to be successful, giving rise to autonomous vehicles, ultra-personalised news feeds and even machines that can beat the best human players in the board game Go.
Amazon is a pioneer in leveraging Machine Learning techniques on their website and in their outbound email content. They can examine the items you search for, what product pages you read and what you buy to accurately determine other things that you might be interested in. While not without its flaws, its success has helped Amazon become the third largest retailer, predominantly without a physical shopping location.
There are clear opportunities in travel to adopt AI in a similar way. Borrowing from the experience of retailers, targeting customers with relevant, personalised offers and services leads to simpler searching and better conversion rates. For example, if we know a traveller’s preferences for a particular type of hotel room, then those could be promoted in search results.
To deliver such personalised services, a travel brand must get to know their customers using three key criteria: time, location and context. These criteria set out when the time is right to engage with a customer, where the customer is at that time (the time might be right, but the customer is not in a position to take action) and how the service on offer relates to what the customer is trying to do. As this interaction will likely differ from person to person, using predictive analytics in this way is becoming an indispensable tool in getting personalisation right.
Starwood Hotels are a great example of a hotel group using predictive data to create a more personalised experience. The more a guest stays at a Starwood Hotel, the more their preferred guest app learns about them. They track their interests over time and use that data to suggest activities and recommendations, and if the app is installed on a guest’s phone, the app experience changes when it senses they have entered one of the hotels to reflect this.
Machine Learning algorithms have a voracious appetite for data. The more data that can be collected, the more accurately they can detect useful patterns. In particular, mobile is a vital source of such data as people carry their phones everywhere and increasingly use it as their primary way to interact with brands. Mobile therefore proves a way to gather information on travellers’ preferences, what they like and need at the various points during their trips.
A number of airlines have started using the abundance of data available, to deliver a more personalised experience. Lufthansa introduced a customised system that analyses data on customer behaviour to determine the best time to offer additional services. Since the system was launched, the number of customers who took an upgrade offer has doubled. United Airlines is another good example, they use a “collect, detect, act” system that analyses 150 variables in a customer profile. Everything from previous purchases to customer priorities is measured in order to present a tailor- made offer. This system has increased United’s year-on-year revenue by more than 15%.
As we move into 2018 and on to 2019, we’ll start to see new entrants specialising in mobile travel analytics enter this space. They will help travel brands leverage this rapidly growing capability by investigating ways to unlock their data and deliver the personalised services that their travellers are now expecting.