When a new technology is hot, the hype can sometimes distort its true meaning and potential. In meeting rooms across the globe, industry evangelists blend notions of big data, business intelligence, data analytics, artificial intelligence (AI), machine learning (ML), and predictive analytics. It may not be their intention to confuse people, but it’s easy to lose track of what predictive analytics is really about.
Predictive analytics versus AI, ML, big data, and data analytics
The improvements in data management technologies over the last five years have led to remarkable advances in data analysis and predictive analytics. With all the associated buzz, people can be forgiven for conflating one workload with another. However, while big data, machine learning, and AI are all related, they also are distinct processes with different objectives and methods. Big data is the management and analysis of large, highly diverse data sets. AI software analyzes data to discover previously hidden insights, like how interest rates may affect the price of a particular stock. And ML software can learn new things from data. You could, for example, show a million digital images of dogs and foxes to an ML tool and it would teach itself to distinguish between the two. Predictive analytics refers to using the analysis of data to make predictions about future events. Typically, this involves assessing data about past events and using algorithms to detect patterns and make predictions. Predictive analytics is, to some extent, a specialized application of big data, AI, and ML. For example, imagine that you operate a fleet of trucks. You have data about engine breakdowns and repairs covering a significant period of time. Big data assembles the engine repair information. AI spots the breakdown patterns. ML teaches the tool to get better at spotting potential breakdowns. Predictive analytics is able to use these inputs to make an educated guess about when an engine is going to fail — and issue an alert about it. This is known as predictive maintenance.
Making it relevant to business travel
Hype is inevitable with new trends like predictive analytics, but it’s helpful to keep a focus on practical and relevant applications of the technology. Of course, you could set up software to analyze a traveler’s complete social graph and seat him on the plane next to someone with the same astrological sign. But, what’s the point in that? A more useful scenario might involve using ML to learn about your company’s booking habits and predict helpful steps to save money or give your travelers a better experience. Let’s say you have to go to Paris on business. If you simply request hotel recommendations on a travel management platform or a consumer site, you will get hundreds of hotel suggestions. You can, of course, dig through them and look at the map, comparing locations and their proximity to the airport and the office. But that’s going to be pretty time-consuming. With predictive analytics, because the booking system knows you, it could recommend hotels where your colleagues usually stay in Paris, allowing you to make decisions based on the choices of people you trust. It can arrange suggestions by distance from your meeting place. Then, it can predict what times of day the taxi routes to your Paris office will be the slowest and most expensive and offer suggestions for when to leave the hotel to get to your meeting on time.
Predictive analytics for travel in small to medium-sized businesses
The use of predictive analytics as a norm in business travel is just around the corner. Egencia for example, has been innovating ways this would support both large and small organizations and the possibilities are exciting. In a large organization, the data analysis underpinning predictive analytics is likely to be contained within the company’s own internal travel history. For small to medium-sized businesses, predictive analytics is probably going to more peer-based considering there is a smaller data set from which to pull learnings. This will provide a great opportunity for those businesses to look to their peers worldwide, and benefit from data learnings from across the globe. Consider the case of a small tech company. They plan to send employees to a major tech trade show in Las Vegas. If they’re working with a travel management company (TMC) that employs predictive analytics, they could get notifications about the best ways to save on fares through advance booking for the event. They can leverage historical travel patterns within the travel management platform to plan the most cost-effective and user-friendly experience for their travelers.
Challenges to developing meaningful predictive analytics in travel
The future of predictive analytics is bright – but traditional TMCs will have to overcome a few obstacles to prepare for a fully functional predictive analytics future. First, there’s the data itself. Right now, the data needed for effective predictive modeling is spread out across numerous online providers and offline services. A single platform model like Egencia's, solves this spread of data by consolidating it all in one place. Then, there’s the machine learning and modeling aspects of the process. Taking a global aspect for example, what customers in one country might consider a good travel experience — or a good deal — might diverge substantially from another. In some regions, the higher-rate luxury hotel is actually an investment in marketing and sales more than a travel expense. In other instances, luxury is strictly prohibited by the travel policy. Any predictive analytics tools must be aware of such nuances and continually get better at learning what they mean. The most serious challenge revolves around making predictive analytics actionable in the travel industry. Egencia has addressed the fact that for the best results, everything must be connected, meaning that the analytics tools must integrate with travel platforms and policy data.
The future of predictive analytics
Right now, TMCs like Egencia are working on ways to make predictive analytics a value-added feature for business travel. It’s a relatively new field, full of fascinating ideas and potential innovations. The challenge lies in perfecting the application of predictive analytics to ensure it helps save money, improves policy compliance, and supports travelers to enjoy a better overall experience.