The passenger lament – “if you have time to spare, go by air” – is as true today as it was in the 1930s. However, the ever-growing capabilities of artificial intelligence and machine learning are corralling Big Data so passengers can game the system to avoid or mitigate travel disruptions. This new technology – of consumer and travel management apps – not only warns about trouble spots at booking but provides escape routes to recover when disruptions occur en route.
“The most exciting development today is delivering information so the traveler doesn’t have to worry, doesn’t have to search and it is all automatic with a few clicks,” OAG chief product officer Vipal Nikum says. “Today, only AI and machine learning can deliver the right information at the right time. What is coming to the forefront now is prescriptive information so passengers can manage disruptions themselves.”
While airlines may blame weather and air traffic control, fully two-thirds of delays and cancellations come from carrier’s operational issues or from mitigating safety risks, according to a 2019 study from Frost & Sullivan. Airlines are now leveraging data to run operations better, which may move the needle on airline reliability as they see on-time performance as a competitive advantage. This, however, is a long-term solution to an increasingly critical problem that costs the passengers billions per year.
So What’s the Problem? “Statistics show 5.6 million departing flights were delayed for an average of 57 minutes, 436,000 flights were cancelled, and the travel plans of over 655 million passengers were disrupted in 2018,” the Frost & Sullivan study reports.
Pilota chief executive officer Saniya Shah estimates disruption costs at $1200 per employee per disruption, making apps like hers important to help passengers take matters into their own hands. Unanticipated travel costs add to the toll, such as non-refundable hotel nights and the fact that when one segment is missed airlines the rest of the itinerary gets cancelled automatically. Furthermore, calculations often don’t include the cost of employee burnout.
“Disruptions usually result in stressed travelers waiting around trapped at an airport, missing meetings, and incurring extra expenses,” says Freebird CEO Ethan Bernstein. “We give travelers the power to immediately find and select new flights, make their meetings and avoid extra nights in hotels, all of which reduces stress.”
Airlines have padded schedules, which boosts their on time numbers, but that’s yielded only a 2.6 percent improvement. Nearly 20 percent of flights still depart or arrive late and, with the projected growth rate in demand – from 4.4 billion to over 8.2 billion passengers flying 47,000 aircraft by 2037 – disruptions will only get worse.
Upping the Score Historically, what help airlines have provided to manage disruption might depend on loyalty or fare class. However now carriers are beginning to look at their operations more holistically, paying attention to Net Promoter Scores – a measurement of how much passengers would recommend the company – and how it relates to on time performance.
“A 1 percent reduction in OTP results in a 0.6 percent reduction in NPS,” according to Karen Miller, GE Aviation director of product for network operations, a Big Data disruption management provider for airlines. “Each point improvement in OTP translates to approximately 5 percent reduction in costs. That may not sound like much, but these are big numbers and airline margins are razor thin,” she explains.
“NPS has been the focus of customer service but about a year ago, the first operations manager talked to me about NPS, trying to understand how OTP contributes to NPS. That’s huge. OTP and NPS should be part of their key performance indicators. The biggest takeaway is that technology allows airlines to protect their revenue, improve OTP and optimize passenger experience.”
Taylor Cornwall, a partner at management consultancy Oliver Wyman, agrees. Cornwall says airlines have often underestimated the cost of disruption, alluding to growing consumer protection laws which have magnified disruption costs in some regions. Such laws are now in 60 countries.
“Today you are seeing an end-to-end review of how to improve performance from schedule design and operational planning to recovery optimization,” Cornwall says. “Customer completion – reaching their destination that day – is an increasingly important metric and many airlines have tools that assist them in better understanding and managing customer impact. Airlines are also improving how they communicate delays and the options for customers to rebook, including advances in self-service platforms.”
The Frost & Sullivan study calculated the cost of delay and cancellation at $33.4 billion in 2018 to airlines alone. “Network disruptions represent the most significant inefficiency cost to the aviation industry,” the report states.
But there’s more; Annakov estimates soft costs – passenger satisfaction and loss of market share – are twice that of such hard costs as the maintenance, crew, refunds and fees airline incur each year, and amounts to 5 percent of global revenues or $35 billion annually.
Factors outside airline control mean we will never get to 100 percent schedule reliability. Even if, as the Frost & Sullivan research suggests, airlines squeeze out the substantial operational challenges costing the industry $74.2 billion – roughly 1.5 times the global industry’s operating profit for 2018 – it would only yield a 2.1 percent improvement in OTP, meaning it will always be up to travelers to remain proactive. Empowering Passengers While airlines work on their side of the equation, passengers and travel managers now have the tools that enable them to take matters into their own hands.
“The general rules of thumb for avoiding delays is earlier is usually better,” Lumo co-founder Bala Chandran advises. “Mainline carriers are typically more reliable than regional carriers. Don't take the last flight of the day. If you're airline agnostic, choose an airline with greater traffic. For instance, if you're flying from Atlanta choose Delta because they're more likely to be able to find another aircraft in case your aircraft is delayed. However, these rules of thumb only go so far, especially if you are already in the middle of a disruption,” Chandran cautions.
Bernstein agrees. “It’s nice to try and avoid disruptions, but what do you actually do when you get disrupted?” he asks. “Usually all you can do is wait – for the airline to give you another option, wait on hold with your TMC to try and find another option, or cancel your meeting. At the end of the day, disruptions are exceptionally difficult to predict and there is no way to prevent them.”
Freebird, App in the Air and Pilota are just some of the emerging technologies used by millions of travelers worldwide. Using flight statistics from OAG, Cirium and FlightStats as their platform foundation, the apps massage historic and real-time data to predict disruptions and then, when they happen, offer solutions to minimize those disruptions. There is also Lumo, an app which powers MTrip, and RoadMap, which was originally designed for travel managers and will soon to be expanded to a consumer app.
“Recovery has always the hardest because changing a flight is complicated,” Chandran says. “It depends on the travel policy, the traveler's status with the airline, the fare restrictions on the original flight. Our TMC partners are able to change these tickets when disruptions arise because they have the delay predictions from us and can then decide what the change fees are.”
As Bernstein explains, Freebird’s approach is direct to the traveler. “What we do is proactively reach out to travelers to provide them with a rebooking link that shows them all available flights across all airlines to get them to their destination. We enable the traveler to book a new flight on any airline, for free, with three taps on their phone in a process that takes 30 seconds or less.”
Nikum notes OAG processes more than 120,000 schedule changes daily. Other data includes origin and destination weather, airports, routes, flight connection variabilities, minimum connection times, visa requirements for hubs that require a visa just to connect, and consumer protection laws in Europe and elsewhere. Then there are such variables as season, airline, time of day, day of week, all needed to estimate delay risk.
“It is all about data, data, data and how you take messy and incomplete raw data and use machine learning and AI to deliver a package of actionable information,” Nikum explains, noting OAG powers such platforms as Freebird and Google.
App in the Air, a personal travel assistant for travel information, boarding passes and the like, was specifically designed for the nearly 40 percent of passengers who do not have managed corporate travel policies. The app is free but premium services, such as real-time flight updates and auto-check in, are also available.
“We wanted to help users make better plans,” says App in the Air CEO Bayram Annakov, adding the app already has six million users accounting for 20 million itineraries worldwide. “Using machine learning and AI to calculate disruptions, we incorporate strategic planning into booking so every flight will display the probability of delays of more than one hour and the probability of a missed connection. Our app learns the preferences of users over time – first by analyzing past flights, then evolving alongside the flier as they interact with the app.”
Pilota CEO Shah sees apps as empowering consumers. “We see this trend across industries with companies like Uber and Seamless/Grubhub,” she says. “We think it is time to apply that to travel. Up until recently data did not have enough precision to make highly confident recommendations for action. With Big Data, AI and newer machine learning methods we can now create that solution for passengers while they are traveling.”