What is Big Data and how does that map to travel?
Gartner defines Big Data as:
“…high-volume, high-velocity and/or high-variety information assets that demand cost-effective, innovative forms of information processing that enable enhanced insight, decision making, and process automation” .
The “Big” in “Big Data” is an ever upward moving target as the volumes of data, the frequency with which this data changes and the number of different types of data captured increases due to technology change, new data inputs and increases in transactional volumes.
From a travel perspective, volumes of data have always been large and will continue to get bigger as the number of online transactions grow. The variety of data captured has also increased significantly. In the past, primary data sets typically associated with travel included fares, inventory and transactions. However we now also consider a multitude of alternative data sources such as search data, conversion data, physical location data, IoT data, traffic and weather. These alternate data sources tend to be more volatile than the core data sets.
Given the massive volumes of data involved, traditional data processing approaches are rarely an option. Instead Big Data style technical approaches such as steaming, sampling, MapReduce among others are required.
There are many potential applications of Big Data techniques in the travel domain. These include:
- Personalisation – based on previous transactions and behaviour, travel companies can predict options that will appeal most to travellers. Amazon is one of the first movers in personalisation with its use of recommendations. Based on a massive number of potential baskets and previous actual checked out baskets, Amazon is able to predict what product is most likely to be added to a user’s current basket. This approach can be adapted in a travel booking concept.
Image source: Amazon.com
- Customer Value – travel companies can categorise travellers based on their customer life-time value. This value can be derived by examining a number of factors including current customer behaviour and previously observed behaviour from customers within a particular segment. For example you may determine that those customers who fly frequently, have actioned full price upgrades in response to a pushed offer and typically arrive just in time for flights are a segment that has a high lifetime value. You may then want to incentivise this segment in some way.
- Fraud detection – by introducing new data sources, fraud detection is able to consider a richer, wider set of data. This allows for a more accurate determination about fraud to be made. An example is Google’s “No CAPTCHA reCAPTCHA” which helps prevent spamming from bots by examining a user’s entire engagement on a site. No Captcha reCaptcha displays like as simple checkbox however Google uses advanced algorithms and a wide set of data (thought to include cookies from other sites, mouse movement across page and IP addresses) to detect whether the user is a bot or a human being.
Image source: Google
- Route planning – it is difficult to calculate the smartest route for travellers, without taking into account real time events and data. Traditional routing algorithms often fell short in this area. However new advancements in big data mean that travel companies can use a combination of live and historical data reported from the mobile devices used by travellers ‘on the go’ to augment route recommendations. For example, systems like Waze can use historical data to identify when a particular route might be busy (e.g. around football stadium for home games) and current data for real time issues (e.g. how long people are at a point to identify when there’s an unexpected traffic jam).
Image source: https://www.waze.com
So Big Data is the answer?
While Big Data approaches are an attractive proposition, they are complicated projects in terms of data gathering and the systems that act on that data. In addition, for a number of reasons beyond the scope of this post, Big Data projects have a very high reported failure rate. In fact, research shows that just over a quarter of organisations say their big data initiatives are a success, according to a Capgemini study .
While Big Data approaches are necessary for gathering certain intelligence, it is important to note that a significant amount of useful data can be gathered more easily by travel companies without needing to adopt a Big Data approach. The term “Small Data” surfaced a few years ago and while it has not gained traction as a term, the concept is interesting. There are lots of readily accessible data points which can help travel companies make intelligent decisions with relatively little cost. Examples include:
- Intelligent defaults based on a user’s previous transactions – if a traveller has searched for Singapore to London the last time they visited your site, this is the most likely search they’ll start with on their next visit.
- Before trying to figure out the best ancillary to promote to a traveller based on a complex myriad of data sources, try promoting the most popular ancillary for that route.
- Based on the user’s location you can pre-populate some of the location specific fields in your site e.g. country, currency, city, phone country code, departure airport.
At Travelport Digital we understand the need to leverage both “Small Data” and “Big Data” insights. By combining this “Small Data” with the “Big Data” to derive detailed insights, we can create offerings and messaging that clearly demonstrate to the traveller that a degree of intelligence was used to make their transaction easier.
If you would like to discuss any of the topics covered in this post or to have a wider discussion about your mobile apps strategy, please contact us here.
 Gartner, February 2016, http://www.gartner.com/it-glossary/big-data
 Information Week, February 2015, http://www.informationweek.com/big-data/big-data-analytics/big-data-success-remains-elusive-study/d/d-id/1318891